Digital PDFs

AA-D113A-TE

March 1980

141 pages

Original

5.9MB

Document:	VAX-11 SORT User's Guide
Order Number:	AA-D113A-TE
Revision:	000
Pages:	141
Original Filename:

OCR Text

VAX-11
SORT

User's Guide
Order No. AA-D113A-TE

January 1979

This manual describes how to use the VAX-11 native mode SORT utility. The
manual is Intended for all users.

VAX-11
SORT

User's Guide
Order No. AA-D113A-TE

SUPERSESSION/UPDATE INFORMATION: This Is a new document for this release.
OPERATING SYSTEM AND VERSION:

VAX/VMX V01.5

SOFTWARE VERSION:

VAX/VMS V01.5

To order additional copies of this document, contact the Software Distribution
Center, Digital Equipment Corporation, Maynard, Massachusetts 01754

digital equipment corporation · maynard, massachusetts

First Printing, January 1979

The information in this document is subject to change without notice and
should not be construed as a commitment by Digital Equipment Corporation.
Digital Equipment Corporation assumes no responsibility for any errors that
may appear in this document.
The softwar·e described in this document is furnished under a license, and
may only be used or copied it in accordance with the terms of such license.
No responsibility is assumed for the use or reliability of software on equipment that is not supplied by DIGITAL or its affiliated companies.

The postage-paid READER'S COMMENTS form on the last page of this
document requests your critical evaluation to assist us in preparing future
documentation.

The following are trademarks of Digital Equipment Corporation:
DEC
DECnet
DECsystem-10
DECSYSTEM-20
DECtape
DECUS
DIBOL
DIGITAL

FOCAL
IAS
MASSBUS
PDP
RSX
UNIBUS
VAX
VMS

7/80-14

Contents
Page

Preface

Chapter 1 Introduction
1.1
1.2
1.3
1.4

Sort Types . . . . .
Input and Output . .
Statistics . . . . . . .
Functions Supported. .

1-1
1-2
1-2
1-2

Chapter 2 Running SORT In Interactive and Batch Mode
2.1
2.2

2.3
2.4

2.5

2.6
2.7

The SORT Command .
Interactive Sessions . '•

2-1
2-5

2.2.1
2.2.2
2.2.3
2.2.4

Batch Sessions.
The SORT Command Description

2-6
2-7
2-12
2-12
2-11·
2-17

2.4.1
2.4.2
2.4.3

2-17
2-20
2-21

A Sample Sort.
Selecting the Sort Type
SORT Statistics .
Samples.

Command Name Qualifiers.
Input-File-Specification Qualifiers
Output-File-Specification Qualifiers.

Specification File

2-23

2.5.1
2.5.2

2-24
2-25

Specification File Records
Specification File Record Formats

Setting up the Keys .
Setting up the Work Files

2-31
2-35

Chapter 3 Calling SORT from User Programs
3.1
3.2
3.3
3.4

3.5
3.6
3.7

File 1/0 Interface . . . . . .
Record 1/0 Interface . . . . .
Programming Considerations .

3-1
3-2
3-2

3.3.1

3-4

Key Comparisons ..

Subroutines (Parameters, Definitions, and Valid Re~urns)

3-5

3.4.1
3.4.2
3.4.3
3.4.4
3.4.5
3.4.6

3-6
3-9
3-11
3-12
3-13
3-14
3-15
3-17
3-20

SOR$INIT_SORT. . . .
SOR$PASS_FILES . . .
SOR$RELEASE_REC.
SOR$SORT_MERGE ..
SOR$RETURN-REC . .
SOR$END_SORT . . . .

Sample MACRO Program . . . . .
Sample COBOL-74/VAX Program . .
Sample FORTRAN IV PLUS Program

iii

Chapter 4 Error Conditions
4.1
4.2
4.3

Command Interpreter Error Messages.
SORT ]~rror Messages .
RMS E:rror Codes . . . . . .

.4-2
4-2

4-8

Chapter 5 Improving SORT Efficiency
5.1

Functional Description. . . .
5.1.1

Sorting Processes . .
5.1.1.1
5.1.1.2
5.1.1.3
5.1.1.4

5-3
5-3
5-5
5-6
5-7

Internal Organization
Buffer Allocation and Work Areas.
Dynamic Memory Usage . .
1/0 Considerations . . . . . . . .

5-10
5-10
5-11

Tuning Procedure . . . . . . . . . . . .

5-11

5.2.1
5.2.2

5-11
5-14

5.1.2
5.1.3
5.1.4
5.1.5
5.2

Record Sort .
Tag Sort . .
Address Sort . .
Index Sort .

5-1
5-3

User Performance Considerations . .
System Manager Performance Considerations .

Glossary
Appendix A Octal/Hexadeclmal/Declmal Conversion
Appendix B Character Set ASCII Collatlng Sequence
Appendix C Data Types
Appendix D Data Structures and Basic Concepts

Index

Figures
2-1
2-2
2-3
2-4
2-5
2-6
2-7
2-8
2-9

SORT's Four Sorting Processes.
File Organization 1/0 Flow . . . . .
Sample Sort Statistics Printout. .
Interactive Session Sample #1
Interactive Session Sample #2 . .
SORT Specification Form . . .
Setting Up the Keys . . . . . . .
Recognizing Data Types and Signed Numbers . .
Specifying Work Files . . . . . . . . . . . . .

2-8
. 2-10
. 2-12
. 2-13'
. 2-15
. 2-25
2-32
. 2-33
2-36

3-1
5-1

5-6

Subroutine Set Summary . . . . . . . . . . . . . . . .
VAX-11 SORT 1\.rchitecture, Main Functional Components
Sample Record Types . . . . . . . . .
Sample Address Sort Output File. .
Sample Index Sort Output File . . .
VAX-11 SORT Operating Phases . . .
SORT Dynamic Memory Usage. .

. 5-10

2-1
2-2
2-3
5-1

SORT Command Summary . . . . . . . . . .
File 1/0 Considerations. . . . . . . . . . . . .
Fixed Position SORT Specification Summary . .
Sorted Output File. . . . . . . . . . . . . . .

. 2-11
. 2-28
. 5-5

5-2
5-3
5-4
5-5

. 3-3

. 5-2
. 5-4
. 5-6
. 5-7
. 5-9

Tables
. 2-3

Commercial Engineering Publications typeset this manual using DIGITAL's
TMS-11 System.
784all

Preface

Intended Audience
This manual is written for the full range of VAXNMS system users, from
beginners to the most advanced level: system operator, applications programmer, system manager, or software developer. Emphasis is on the how to
use information, and detailed descriptions of SORT internals are kept to a
minimum.
You can use SORT as an interactive utility (Chapter 2), or as a set of
subroutines, callable from VAX-11 programming languages (Chapter 3).
New users or those with simple sort requirements, can learn how to use SORT
by reading Chapters 1and2. To use SORT efficiently or for more than simple
sorts, read also Chapters 3 and 5.

vii

Structure of this Document
Chapter 1 introduces the VAX-11 SORT program and describes its environment, features, and requirements, and explains user requisites.
Chapter 2 explains how the SORT command is used to run VAX-11 SORT
interactively or in batch mode.
·
Chapter 3 explains how to call SORT routines from user programs, and describes how to use subroutine parameters.
Chapter 4 provides complete lists of SORT error messages and recovery procedures.
Chapter 5 provides information and programming techniques for improving
VAX-11 SORT efficiency.

The Glossary defines terms used in this manual.
Appendixes A, B and C consist of helpful programming aids such as: code
conversion charts, character sets, and data types used by VAX-11 SORT.
Appendix D summarizes basic concepts.

Finally, page references to key terms appear in the index.

Associated Documents
The following documents are relevant to VAX-11 SORT users:
• PDP-11 SORT Reference Manual
• VAXNMS Primer
• VAXNMS Summary Description
• VAXNMS Command Language User's Guide
• VAXNMS System Messages and Recovery Procedures Manual
• Introduction to VAX-11 Record Management Service
• VAX-11 Record Management Services User's Guide
• VAX-11 Record Management Services Reference Manual
• VAX-11 Software Installation Guide
• VAXNMS System Services Reference Manual
• VAX-11 Common Run-Time Procedure Library Reference Manual
• VAX-11/780 Architecture Handbook
• VAX-11/780 Processor Handbook
• VAX-11/780 Software Handbook
• VAX-11/780 Technical Summary

Symbology
You will encounter the following symbols, colors, and special graphics in this
manual.
dollar sign The system prompt; indicates that the VAXNMS com$ mand interpreter is ready for command input. The next
$ prompt indicates successful completion of command
processing, and the system's readiness to accept another command.
In addition, the $ must appear in the first character
position of a command to be executed in an indirect
command file.
Return

Indicates RETURN or ESC key entry required. Pressing this key after entering a full command line ends the
command input and begins processing.
When using the prompted command format, ~ or
is required after each command segment.

Square Brackets

[]

Used in manual text to indicate qualifiers; not entered.
Used in command syntax to indicate enclosed portion is
optional.

Braces

Used in manual text to indicate input options where

Used in this manual text to indicate variable data input
(typically some number value); not entered.

n one in the vertical list must be selected; not entered.

Underscore

Indicates an entered underscore character.

Hyphen

Indicates line continuation.

Comma

Commas are entered to separate listed subqualifiers.

Circumflex

Represents the CTRL key on many terminals. Normally entered simultaneously with the .alphabetic character that immediately follows.
For example, AC is the same as CTRL/C.

Uppercase Letters
ABC

Indicates command inputs that must be entered as
shown.

Lowercase Letters
abc

Used in text to describe the command syntax; not
entered.

Red print

Indicates characters you type at the terminal. All system printouts appear in black print.

Shading

Used to highlight that portion of an example that is
being described in text.

Chapter 1
Introduction
VAX-11 SORT rearranges and reformats records in any VAX-11 record man"'.
agement service (VAX-11 RMS) file organization. SORT consists of two functional parts: a control program called the utility, and a callable subroutine
package. The utility can be used in an interactive terminal session or in batch
mode using the VAXNMS DIGITAL command language (DCL) SORT command. The callable subroutines are invoked by the SORT utility. Users can
write control programs in most VAX-11 languages using these callable
subroutines.
You can invoke SORT interactively by entering a SORT command with qualifiers and input/output parameters. The command specifies one of four sort
types and the sorting keys. During program execution, SORT indicates all
errors. At the completion of each session, SORT prints a statistical summary.

1.1 Sort Types
The four sort types (or sorting processes) are:
• Record Sort
• Tag Sort
• Address Sort
• Index Sort
Record Sort produces a reordered data file sorted by specified key fields (that
is, entire records are reordered). This sort uses any VAXNMS input device
and can process any valid VAX-11 RMS format. Record, a relatively slow
sort, is the default process.
Tag Sort produces the same kind of output file as record sort by sorting only
the record keys. Tag sort then randomly reaccesses the input file to create a
resequenced output file according to those record keys. This method conserves
temporary storage, but can only accept input files residing on disk.

1-1

Addre11 Sort produces an address file. That is, a reordered address file, on

disk only, of record's file addresses (RFAs). The address file, sorted by record
keys, can be used by programs as an index to read the original file in the
desired sequence. This is the fastest of the four sorting processes.
Index Sort produces an address file containing the key field of each data record
and a pointer (RFA) to its location in the input file. The address file can be
used by programs to randomly access data from the original file in the desired
sequence. Like address sort, this is a high-speed process.

For more information on sort types, see Chapter 2.

1.2 l·nput and Output
As input, SORT accepts sequential, relative or indexed-sequential data files
containing records of fixed, variable, or variable with fixed-length control
(VFC) format. Character, binary, or decimal data types, and files from disk,
magnetic tape, card reader or terminals are accepted.
As output, SORT produces sequential, relative or indexed-sequential data
files. These files can be of fixed, variable or VFC format and output to disk,
magnetic tape, printer or terminal. In addition, SORT outputs address files
(on disk only) for sequential access by programs.

1.3 Statistics
SORT prints statistics at the end of each session. These statistics include:
• Elapsed execution time
• Number of records read, sorted, and output
• The longest record length
For more information on statistics and how they can be useful, see Section
2.2.3 and Chapter 5.

1.4 Functions Supported by VAX-11 SORT
1. Sort types: record, tag, address, index.

2. File organizations as input and output: sequential, relative, indexedsequential. All VAX-11 RMS file types are supported.
3. Record format for input and output: fixed, variable, and VFC. All
VAX-11 RMS record formats are supported.
4. All VAXNMS devices are supported for input and output.

5. Multivolume support as provided by VAX-11 RMS.

1-2

Introduction

6. Callable subroutine package. VAX-11 programming languages producing
native mode code are supported. Included are:
VAX-11 COBOL-74
VAX-11 FORTRAN IV-PLUS
VAX-11 MACRO
VAX-11 BLISS
7. Controlled by command string or specification file.
8. Free field and fixed position specification file formats.
9. Data Types:
• Character data is ASCII representation
• Binary data is VAX representation
• Packed decimal data is VAX representation
• Zoned data is VAX representation
• Decimal data supports:
-

leading separate sign
leading overpunched sign
trailing separate sign
trailing overpunched sign

10. Ascending/descending output based on each key field.
11. Output file blocking and allocation size.
12. Sort statistics provided at completion.
13. ASCII collating sequence for character keys.
14. RSX SORT-11 utility option.

Introduction

1-3

Chapter 2
Running SORT in Interactive and Batch Mode

This Chapter explains how to use the SORT command to sort files interactively or in batch mode.

2.1 The SORT Command
The SORT command consists of three parts: the command name, the input
file specification parameter, and the output file specification parameter, in
that order. Each part must be separated by one or more spaces or tabs, and is
invoked by terminating with ~ when the command is entered as a continuous command string.
This section describes how sorts are performed using the SORT command
without the specification file qualifier. The specification file is a more sophisticated method of controlling SORT, and therefore is described later in Section 2.5. The specification file is a command qualifier and should not be
confused with the file specifications for the input and output files.
Format:
$

40 RT C u a 1 if i er s ~~n Put - f i 1 e - s Pe c if i cation C u a 1 i f i e rs}..
ci

'\: u t Put - f i 1 e - s Pe c i f i c at i on Cci u a 1 i f i e r s J/([IT)

Command Name (SORT)

•

SORT is the command name· that invokes the VAX-11 SORT utility. Command name qualifiers specify the sort process, describe the sorting key(s),
specify the number of work files, indicate the specification file if a sort other
than a standard sort is to be performed, and finally indicate whether the
VAX-11 SORT utility or the RSX SORT-11 utility is to be invoked.

2-1

Input Fiie Specification Parameter

This parameter specifies the physical location of the input file (see Appendix
D for additional file specification information). Input file qualifiers define the
input file attributes such as record format and file size.

Output F'lle Specification Parameter

This parameter specifies the physical location of the sorted output file (see
Appendix D for additional file specification information).
Output file qualifiers define the output file attributes such as record format,
record size, block size, file organization, allocation quantity, contiguous allocation, overlay existing file, and bucket size.
The VAXNMS command interpreter will prompt you for input and output
file specifications if they are not entered in the first command string. The
following is an example of prompted format:
$.._J...sw.o....R....,r..1.....K
...
'E
••'._'=
.............................,.......=..,. , SIZE= 80) (Bfil) .....i---command with qualifiers
$_Fi 1 e :
R 1 0 0 SQ (8If)
input-file-specification
$ _ 0 u t Put :
TEST • TM P (8(f)
output-file-specification

The following example shows how the SORT command is structured:
SORT Command

rcommand
Name
f$

- '.

Command Parameters

. . . . ""j'.
'~ET
Ke1wort
input- f.l 1 e - s P e [ u t - f i. 1 e-sPec
·i=i

Command Qualifiers
(/keyword eyword/keyword)

File Qualiliersj
(/keyword/keyword/keyword)

,.___ _ _ _ _ Subquallflers

(=n) or (=(keywordfeyword))

Values
(=n)

Notes: 1. keywords may be truncated and are unique.
2. n indicates variable data input (typically some number value).

Table 2-1 summarizes all the SORT command qualifiers, subqualifiers, and
input values. The complete details on qualifiers and input values are discussed in Section 2.4.

2-2

Running SORT in Interactive and Batch Mode

Table 2-1: SORT Command Summary
Notation Used:
• Underlined upper-case characters indicates the minimum entry required.
• Brackets [ ] indicate enclosed portion is optional.
• If several enclosed words are listed vertically, only one may be used.
• Qualifiers, subqualifiers and values that must be specified are shown without brackets.
• Braces I I indicate a selection must be made from the vertical list.
• Defaults are shown in bold type.

Subqualifiers
and Values

Command
Qualifiers

Notes

$SORT

RECORDl~
r'ROCESS~ [~gRESS
INDEX

/KEY=

/KEY: is not required if specified in a specification file.

( [NUMBER=[l-101]
,POSITION=ll-163831
1-255 for CHARACTER data type}
,SIZE= l, 2, or 4 for BINARY data type
{ 1-31 for DECIMAL data type
CHARACTER

]

.~IN ARY

,iONED
[ ,D.E.QIMAL
,PACKED_DECIMAL

TRAILING_SIGN is default if data type
is DECIMAL.

,LEADING_SIGN
[ ,TRAILING_SIG~

,,SEPARATE_SIGN
[ ,OVERPUNCHED_SIGNj

OVERPUNCHED_SIGN
data type is DECIMAL.

is default if

,ASCENDING)
[ ,DESCENDING )j
[!WORK_FILES=[0,2-101]

[!SPECIFICATION [=file-specification[)

SYS$INPUT is default file name.

[!RSX11]

VAX-11 SORT is default. /RSXll requires
PDP-11 SORT command switches. Refer
to the PDP-11 SORT Reference Manual.
(continued on next page)

Running SORT in Interactive and Batch Mode

2-3

Table 2-1: SORT Command Summary (continued)
Input File
Qualifiers

Subqualifiers
and Values

Notes

input-file-specification

[!.EORMAT=

See Appendix D for file-specifications.
DAT is default file-type.

(RECORD_-8IZE=[l-16383LJ

GEILE_SIZE= (1-4294967295])

Output File
Qualifiers

RECORD_SIZE
specified.

Subqualifiers
and Values

output-file-specification

[!.EORMAT] =

not

normally

FILE_SIZE is not normally specified.

Notes
See Appendix D for file specifications. The
default output file type is the same as input file type.

FIXED=[l-16383]
]
(VARIABLE:=[l-16383]
(CONTROLLED=[l -16383]

FIXED record format is default if sort process is index or address.

GSIZE=[l-25fil]

Used for VFC records only. Default value is
2 if CONTROLLED is specified and SIZE
is not.

(,BLOCK_SIZE= (18-327671 )]

For magnetic tape files only.

/SEQUENTIAL
]
/RELATIVE
[ /INDEXED_SEQUENTIAL

Default is the input file organization if sort
process is record or tag, otherwise
/SEQUENTIAL
is
default.
If
/INDEXED_SEQUENTIAL is specified,
/OVERLAY must be specified.

[I,ALLOCATION=(1-4294967295 J]

Required if /CONTIGUOUS is specified.
The default value is determined by the
number of records sorted.

[!QONTIGUOUS]

/NOCONTIGUOUS is default.
/CONTIGUOUS
is
invalid
/ALLOCATION is not specified.

[!Q.YERLA Y]

/NOOVERLA Y is default. /OVERLAY is
required if /INDEXED SEQUENTIAL output file organization is specified.

[/fil.ICKET_SIZE=[l-321]

Default value is the same as the
input file value if the input and output file
organizations are the same, otherwise
default is 1.

2-4

Running SORT in Interactive and Batch Mode

2.2 Interactive Sessions
To invoke SORT in interactive mode simply enter the s·ORT command. Any
errors in the command are immediately reported at your terminal (see Chapter 4, Error Conditions). At the end of a successful run, SORT prints the
statistics message (see Section 2.2.3).
SORT accepts two kinds of command formats: a keyboard-oriented command
string containing all the command qualifiers (excluding /SPECIFICATION),
or a keyboard-oriented command string containing the /SPECIFICATION
qualifier pointing to a specification file containing the command qualifiers.
For example:
$

SORT/KEY=<POSITION=l ,SIZE=10) inPut-file-sPecification
outp1.1t-f i le-specification ffiJ

or:
$

SORT/SPECIFICATION=file-specification
inPut-file-sPecification output-file-specification l8@

The use of the specification file is the more involved method and therefore
explained in Section 2.5.
In order to specify a sorting sequence, you must select key fields within the
data itself. Remember, SORT reorders the entire file. The information provided in Section 2.6 can help you to set up the key fields (keys).
You can extract key information from a file and store it in a reordered format
for future use in accessing data in your original file in the order of your
reordered file. In addition, the contents of your sorted file can be entire
records, key fields with record pointers, or record indices relative to the position of each record within the file. Your intentions for the sorted output file
usage, together with input and output file organizations, determine what sort
process to use. The information provided in Section 2.2.2 can help you to
choose the correct sorting process.
Because SORT is designed to process all RMS file organizations, you also
must consider how to direct the sorting process you have chosen, so that your
output file organization will be usable on your peripheral device. The information provided in Section 2.2.2 and Table 2-2 compares file organizations and
sorting processes.
If your sorting task requires more than two work files, Section 2.7 can help you
to set up additional work files. Most sorts will normally use the default number of work files.

Finally, you must specify input and output file specifications. Appendix D
reviews the standard VAX/VMS file-specification information, and file specification qualifiers are summarized in Table 2-1, and described in detail in
Section 2.4.

Running SORT in Interactive and Batch Mode

2-5

2.2.1 A Sample Sort
Users can invoke the SORT command by simply providing the required key
position and size for a single key and the file name of a single input file
located on the user's default disk.
The format of the command is:
$

SORT/KEY=<POSITION=Cl-18383] tSIZE=C1-255J)
inPut-file-sPecification output-file-specification IBEf)

This means:
• A record sorting process is performed on the specified input file
• The input file key data type must be character
• The input file must reside on the user's default disk
• All the records in the input file are reordered in the output file in ascending
alphabetic order
• Input file type DAT is assigned, and output file type DAT is assigned
• SORT assigns two work files for temporary storage
• Output file organization is the same as the input file organization
• Output file record format is the same as input records format
• Output file bucket size is the same as input file bucket size
• SORT statistics are printed at the terminal that executed the sort
An abbreviated representation of the preceding command example is:
$

SORT/K=<P0=1 tSI=80) INPUT OUTPUT

CB:IT)

Description:
If you specify the key position and size, and character data type by default;

this sort reads the single input file specified (on the user's default disk), sets
up two work files on the user's default disk, and performs a record sort.
This process creates an output file named OUTPUT.DAT having the same
file organization as the input file. All the records in INPUT.DAT are reordered in ascending alphabetic order in the output file. The alphabetic order is
determined by the contents of the 80-character key field (SI=80) starting in
position one (PO=l) of each record.
NOTE:
A quick test can be run at your terminal by using
SYS$0UTPUT as the output-file-specification. This technique
displays the sorted output file before the sort statistics are
printed.

2-6

Running SORT in Interactive and Batch Mode

Finally, upon completion of the run the sort statistics are printed at the
terminal that executed the sort.

2.2.2 Selecting the Sort Type
SORT offers a choice of four processes: record, tag, address, and index. You
specify the sort process by using the proper qualifier in the command or in the
specification file code. Each process has its particular input requirements,
processing methods, device requirements, and resultant output files.
SORT provides four sorting techniques:
• RECORD (/PROCESS=RECORD, or SORTR if specification file)

Record sort produces a reordered data file sorted by _specified key fields
(that is, entire records are reordered). This sort can be used on any acceptable input device, and can process any valid VAX-11 RMS format. Record,
a relatively slow sort, is the default process.
• TAG (/PROCESS=TAG, or SORTT if specification file)

Tag sort produces the same kind of reordered data file as record sort by
sorting only the record keys. This method conserves temporary storage, but
can only accept input files residing on disk. Tag sort is faster than record
sort, if the key size is much smaller than the record size and the file size is
small so that the reaccessing process is short.
• ADDRESS (/PROCESS=ADDRESS, or SORTA if specification file)

Address sort produces an address file without reordering the input file. That
is, a reordered address file (on disk only) of record's file addresses (RF As).
The address file, sorted by record keys, can later be used as an index* to
read the original file in the desired sequence. Any number of address files
may be created for the same data base. A customer master file, for instance,
may be referenced by either customer-number index or sales-territory index
for different reports. This is the fastest of the four sorting processes.
• INDEX (/PROCESS=INDEX, or SORTI if specification file)

Index sort produces an address file containing the key field of each data
record and a pointer (RFA) to its location in the input file. The address file
can be used by programs to randomly access data from the original file in
the desired sequence. Like Address sort, this is a high speed process.

Figure 2-1 summarizes these options to help you determine which process is
best for your sorting application. Chapter 5 provides additional information
regarding sorting processes where performance considerations are important.

* Not indexed by VAX-11 RMS.

Running SORT in Interactive and Batch Mode

2-7

Figure 2-1: SORT's Four Sorting Processes
Input Data File

}·~\....

default process.
(Slowest Process)

(Dlsk,Magtape,Cards,Termlnal)

~ of Reordered
Output Data Fiie
Rec.ords

Entire
Records.

Sorted by Keys.

D_ __

--~~---~-~~-

Temporary
Storage
(Work Files)

Disk only,
2-10 Files.

C _ __
B
A--(Disk,ANSI Magtape,
Printer, Terminal)

(Faster than Record sort
if Key Size is smaller than
record and file size is small.)
Uses less temporary
storage than Record sort.

Input Data File

A_ __

(Fastest
Process)
Uses minimum
temporary
storage.

8 _ __

c ___
D
(Disk Only)

Output Address Fiie
of Reordered Record's Fiie
Address (RFA) Records
(fixed 6-byte records).
RFA in binary
RFA in binary
RFA in binary
RFA in binar
(Disk Only)
For !ater use to access the original
File when this particular sequence
is desired.

Temporary
Storage
(Work Flies)

INDEX Sort

Disk only,
2-10 Flies.

Output Address Fiie
of Record's Fiie Address (RFA)
Records (fixed 6-byte records)
plus Key Fields.
RF in nary + ey
RFA in binary + Key
_ _ _ _ _.....,.RFA in binary + Key

(A fast process if Key
Size is less than Record.)

F-MK-00018-00

2-8

Running SORT in Interactive and Batch Mode

RFA in binary + Key
(Disk or ANSI Magtape)
For later use to randomly access the
original File in the desired sequence.

Fiie 1/0 Co11slderatlons

Input and output file organizations are another important factor in determining which eort type to use. Figure 2-2 shows how the 1/0 flows through SORT,
and Table 2-2 list all possible 1/0 combinations and shows the default output
file organb:ations.
Inputs to v,,X-11 SORT can be files of sequential, relative, or indexed organization containing records of fixed, variable, or VFC format from disk, magnetic
tape, card reader, or terminals.

Input parameters to the sort program are either provided by RMS after processing the input file header records, or specified in the command in the form
of input-file-specification qualifiers (that is, /FORMAT ... ).
Outputs frc1m VAX-11 SORT are files of records reordered by key fields and are
created in sequential, relative, or indexed organization. 'rhese files may contain record types of fixed, variable, or VFC format. Output files can be written
to disk, m.ignetic tape, printer, or terminals.

Sorted output address files of 6-byte RF As in binary coded records are output
to disk only for sequential access by programs. These output address files are
intended {iJr software use as indices into input files, and cannot be output to

printers or terminals without further processing.
Output parameters to the sort program are specified in the command in the
form of output-file-specification qualifiers (that is, /FORMAT ... ).

Running SORT in Interactive and Batch Mode

2-9

Figure 2-2: File Organization 1/0 Flow

INPUT

OUTPUT
SEQUENTIAL
DATA FILE

SEQUENTIAL
DATA FILE

SORT

SEQUENTIAL
ADDRESS FILE

RELATIVE
DATA FILE
RELATIVE
DATA FILE

INDEXED-SEQUENTIAL
DATA FILE

SEQUENTIAL
ADDRESS 'FILE

INDEXED
DATA FILE

SEQUENTIAL
ADDRESS FILE

NOTES:
1. RECORD & TAG produce reordered data files of the same organization as input
by default.
2. INDEX produces reordered address files of RFA's plus keys in sequential file
organization.
3. ADDRESS produces reordered address files of RFA's only in sequential file
organization.

F-MK-00019-00

2-10

Running SORT in Interactive and Batch Mode

Table 2-2: File 1/0 Considerations

Type of
Input File

Sequential
Data File

Relative
Data File

IndexedSequential
Data File

Sort
Process

Output File
Organization
Specified

Results

Record

Sequential
Relative
Indexed-Seq

Tag

Sequential }
Relative
Indexed-Seq

Same as for record.

Address

Sequential }
Relative
Indexed-Seq

Sequential address file of RFAs.

Index

Sequential }
Relative
Indexed-Seq

Sequential address file of RFAs with keys.

Record

Sequential
Relative
Indexed-Seq

Tag

Sequential
Relative
Indexed-Seq

Address

Sequential
Relative
Indexed-Seq

Index

Sequential
Relative
Indexed-Seq

Record

Sequential
Relative
Indexed-Seq

Tag

Sequential
Relative
Indexed-Seq

Reordered sequential data file.
Reordered file of data records.
Populates (overlays) an already existing IndexedSequentiai output file with reordered data records.

Same as above for each process.

Sequential--~

Address

Relative
Indexed-Seq
~____,

Index

Sequential
Relative
Indexed-Seq
....,.,..,

Note: The default output file organization is shown in italic type.

Running SORT in Interactive and Batch Mode

2-11

. 2.2.3 SORT Statistics
Statistics are automatically printed at the completion of each sort session.
These consist of: elapsed execution time, the number of records read, sorted,
and output; the longest record length; the multiblock count used and the
multibuffer count used for input and output; the merge order; the number of
merge passes; the working set size used; the number of initial runs; and the
virtual memory used for the sort tree.
In addition, SORT statistics include statistics kept by VAX/VMS for the
number of buffered and direct 1/0 operations, CPU time, and the number of
page faults. Figure 2-3 illustrates a typical SORT statistics printout of a
single sequential input file (filename RlOOSQ.DAT) that is 10,000 records in
length, and each record is 80 characters long. The sorting is done on an 80
character key starting at position 1 of each record. The output filename is
TEST.TMP and is output in the same format as the input file by default.
The command string that caused the sample printout in Figure 2-3 was:
SORT/KEY=CPOS=1,SIZE=80) C8fll
$_File: R100SQ (8fll
$_01.ttP1.tt:TEST.TMP ~
$

The statistics can be used to help tune the parameters you specify for a
specific sort, such as the best working set quota size to use (see Chapter 5,
Section 5.2.2.5).

Figure 2-3: Sample Sort Statistics Printout
SORT STATISTICS:
RECORDS READ:
10000
RECORDS SORTED: 10000
RECORDS OUTPUT: 10000
MM{ I MUM WORK I NG SET USED: 128 0
VIRTUAL MEMORY ADDED: 2380328
DIRECT IO COUNT: 227
BUFFERED IO COUNT: 23
PAGE FAULTS: 15598
ELAPSED TIME: 00: 02: 28. 97 C)

LONGEST RECORD LENGTH: 80
INPUT MULTI BLOCK COUNT: 11
OUTPUT MULTI BLOCK COUNT: 20
INPUT MULTI BUFFER COUNT: 2
OUTPUT MULTI BUFFER COUNT: 2
NUMBER OF INITIAL RUNS: 38
ORDER OF THE MERGE:
7
NUMBER OF MERGE PASSES: 2
CPU TI ME: 5055 8

Notes:

O Maximum working set used is in blocks.

fJ Virtual memory added is in bytes.
C) Elapsed time is the total sort run time from start to end in hrs: min: sec.

1/100secs.

8 CPU time is the data processing time less 1/0 time in 1/1 OOsecs. (that is, 5055
is 50 seconds and 55/100th's seconds.)

2.2.4 Samples
Figure 2-4 shows a step-by-step session for an interactive sort on a single key.
Figure 2-5 shows how an interactive sort would appear when sorting on two
keys.

2-12

Running SORT in Interactive and Batch Mode

Figure 2-4: Interactive Session Sample #1

Step 1: Observe the Input file you want to sort to determine where the key fields are
located and their size.
To sort this Input file named BOATS.LST In alphabetic order by manufacturer you must
specify a single key field starting at character position 2 and having a key field size of 1O
characters.

MANUFACTURER

MODEL

RIG

LENGTH

NORTHERN
CHALLENGER
OLYMPIC
EASTWARD
AMERICAN
LINDSEY
WINDPOWER
CAPE DOf~Y
VENTURE
SALT
AMERICAN
HUNTER
TANZER
ALBIN
GRAMPIAN
CHRIS-CRAF
ISLANDER
COLUMBIA

37
al
ADIJENTURE
HO
28-MS
39
IMPULSE
TYPHOON
222
19
28
27
28
BALLAD

KETCH
KETCH
KETCH
M/S
M/S
M/S
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP

Z-3a

CARIBBEAN
36
41

az
2a

28
39
16
18

25
28
27
28
30

35
36
41

WEIGHT BEAM

PRICE

1a1000
281700
2a12so
7 1000
5t500
1a15oo
650
1 t 900
21000
21600
a1000
G t500
6t800
71276
11 t800
18t000
13t450
20 t 700

$5(11000
$511228
$80 I 500
$15 1900
$18t895
$351800
$3t500
$4t285
$3t56ll
$Gt580
$91895
$1a1sss
$17t500
$27t500
$28t675
$37t850
$31 t730
$ll8 tll90

11
13
13
09
08
12
07

08
07
07
08
08

10
10
10
11
11
11

/KEY =(POSITION=2,SIZE= 10)

Step 2: Enter the following SORT command to sort the input file named BOATS. LST and
create an output flle named BOATS.ALB:

SORT/KEY=<POS=2tSIZE=10)

BOATS+LST BOATS.ALB (5£!)

Running SORT in Interactive and Batch Mode

2-13

Step 3: Observe this printout when SORT has completed.
SORT STATISTICS:
RECORDS READ:
18
RECORDS SORTED: 18
RECORDS OUTPUT: 18
MAXIMUM WORKING SET USED: 200
VIRTUAL MEMORY ADDED: 404982
DIRECT IO COUNT: 2
BUFFERED I 0 COUNT: 17
PAGE FAUL TS:
146
ELAPSED TIME: OO:OO:Oi.87

LONGEST RECORD LENGTH: 57
INPUT MULTI BLOCK COUNT: 20
OUTPUT MULTI BLOCK COUNT: 32
INPUT MULTI BUFFER COUNT: 2
OUTPUT MULTI BUFFER COUNT: 2
NUMBER OF INITIAL RUNS: 0
ORDER OF THE MERGE:
0
NUMBER OF MERGE PASSES: 0
CPU TIME: 54

Step 4: Examine your newly sorted output file named BOATS.ALB. Notice that the records are now in alphabetical order.

2-14

MANUFACTURER

MODEL

RIG

LENGTH

ALBIN
AMER I CAN
AMERICAN
CAPE DORY
CHALL ENGEi~
CHRIS-CRAF
COLUMBIA
EASTWARD
GRAMPIAN
HUNTER
ISLANDER
LI ND SEY
NORTHERN
OLYMPIC
SALT
TANZER
VENTURE
WIND POWER

BALLAD
26
26-MS
TYPHOON
41
CARIBBEAN
ll 1
HO
2-34
27
38
38
37
ADVENTURE
18
28
222
IMPULSE

SLOOP
SLOOP
M/S
SLOOP
KETCH
SLOOP
SLOOP
M/S
SLOOP
SLOOP
SLOOP
MIS
KETCH
KETCH
SLOOP
SLOOP
SLOOP
SLOOP

30
26
26
18
ll 1
35
41
2l!
3l!
27
36
38
37
42
25
28
22
18

Running SORT in Interactive and Batch Mode

WEIGHT BEAM
7t27G
t.!1000
5,500
1 ,900
261700
181000
20,700
71000
111800
8t500
131l!50
1l! t500
1l!1000
2llt250
21600
6, 800
21000
650

10
08
08
06
13
11
11
08
10
09
11
12
11
13
07
1(I
07
07

Pl~

I CE

$27 1500
$81885
$181885
$41285
$51 1228
$371850
$l!8 1ll80
$15t800
$2B1G75
$ ltJ 1888
$31 ,730
$351800
$501000
$801500
$81580
$171500
$3 15Gl!
$31500

Figure 2-5: · Interactive Session Sample #2

Step 1: Observe the Input flle you want to sort to determine where the key fields are
located and their size.
To sort this Input file named BOATS.LST In ASCII alphanumeric order first by beam, and
then by price, you must specify two keys. The first key (or primary key) field starts at
character position 47 and has a size of 2. The second key starts at character position 51
and has a size of 7.

MANUFACTURER

MODEL

RIG

LENGTH

NORTHERN
CHALLENGER
OLYMPIC
EASTWARD
AMERICAN
LINDSEY
WINDPOWER
CAPE DORY
VENTURE
SALT
AMERICAN
HUNTER
TANZER
ALBIN
GRAMPIAN
CHRIS-CRAF
ISLANDER
COLUMBIA

37
ll 1
ADlJENTURE
HO
28-MS
39
IMPULSE
TYPHOON
222
19
28
27
28
BALLAD
2-3ll
CARIBBEAN
36
ll1

KETCH
KETCH
KETCH
M/S
MIS
MIS
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP

37
ll 1
ll2
2ll
28
39
18
18
22
·25
28
27
28
30'
3ll
35
38
ll1

WEIGHT BEAM

PRICE

1llt000
2Gt700
2llt250
7 , 000
5t500
1llt500
850
1 t800
ZtOOO
2 t GOO
lltOOO
Gt500
Gt800
7t278
11 tBOO
18t000
13tll50
20t700

$50t000
$51 t228
$80t500
$15,900
$18t895
$35t800
$3t500
$llt285
$3t5Gll
$6t580
$8t885
$1llt889
$17t500
$27t500
$29t875
$37t850
$31 t730
$ll8tll90

11
13
13
09
08
12
07
OB
07
07
08
09
10
10
10
11
11
11

/KEY=(POS=47,~2)

/KEY=(P0S=51,SIZE=7)

Step 2: Enter the following SORT command to sort the input file named BOATS.LST and
create an output file named BOATS.BEM:

SORT /KEY= ( POS=tn 1SIZE=2) /KEY= ( POS=51 t!3I:ZE=7)
BOATS+LST BOATS+BEM (Bffl

Running SORT in Interactive and Batch Mode

2-15

Step 3: Observe this printout when SORT has completed.

SORT STATISTICS:
RECORDS READ:
18
RECORDS SORTED: 18
RECORDS OUTPUT: 18
MAXIMUM WORKING SET USED: 200
VIRTUAL MEMORY ADDED: 202240
DIRECT ID COUNT: 2
BUFFERED IO COUNT: 17
PAGE FAUL TS: 142
ELAPSED TIME: 00:00:01.85

LONGEST RECORD LENGTH: 57
INPUT MULTI BLOCK COUNT: 20
OUTPUT MULTI BLOCK COUNT: 17
INPUT MULTI BUFFER COUNT: 2
OUTPUT MULTI BUFFER COUNT: 2
NUMBER OF INITIAL RUNS: 0
ORDER OF THE MERGE:
0
NUMBER OF MERGE PASSES: 0
CPU TIME: as

Step 4: Examine your newly sorted output file named BOATS.BEM. Notice that the
records are now In order first by beam width, and second by price.

MANUFACTURER

MODEL

RIG

LENGTH

CAPE DORY
WINO POWER
VENTURE
SALT
AMERICAN
AMERICAN
HUNTER
EASTWARD
TANZER
ALBIN
GRAMPIAN
ISLANDER
CHRIS-CRAF

TYPHOON
IMPULSE
222
18
26
28-MS
27
HO
28
BALLAD
2-3ll
36
CARIBBEAN
ll1
37
38
41
ADVENTURE

SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
M/S
SLOOP
MIS
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
SLOOP
KETCH
M/S
KETCH
KETCH

18
18
22
25
26
26
27
24
28
30
3ll
38
35
ll1
37
39
ll1
ll2

COLUM~IA

NORTHERN
LINDSEY
CHALLENGER
OLYMPIC

2-16

Running SORT in Interactive and Batch Mode

WEIGHT BEAM

PRICE

1 ,900
850
2,000
2,900

$4,285
$3,500
$3,584
$8,580
$8,885
$18,885
$14,999
$15,800
$17 ,500
$27,SOO
$29,675
$31 ,730
$37,850
$ll8,ll90
$50t000
$35,800
$51 ,220
$8(1,500

a,ooo

5,500
9,500
7,000

s,soo
1,218

11 , 800
13,450
10,000
20,700
1a,ooo
1a,500
28,700
za,250

08
07
07
07
08
08
09
09
10
10
10
11
11
11
11
12
13
13

2.3 Batch Sessions
To run the same sort as shown in Figure 2-4 using batch mode, perform the
following steps:
Step 1: Create a command file named BOATSl.COM as follows:
$
$

PRINT BOATS.LST
SORT/KEY=<POS=2tSIZE=10) BOATS.LST BOATS.ALB
PRINT BOATS.ALB

Step 2: Enter this command:
$

SUBMIT BOATS1.COM

filj)

Observe this response:
Job n entered on queue SYS$6ATCH

Step 3: Observe that the input file, output file, sort statistics, and batch
statistics are all printed on the system printer.

2.4 The SORT Command Description
NOTE:
Review the SYMBOLOGY in the front of this manual before
continuing.
Format:
$

SORT[qualifiersJ inPut-file-sPecification[qualifiersJ
outPut-file-sPecification[qualifiersJ ~

2.4.1 Command Name Quallflers
Abbreviated Example:
$

s 0 Rrl/Jp::~::a::e::fi::~::s::;::ri::z::k:iiNt:!B\fJ%M\\~fdBJiliX:x~::li.~~ltt~il1S\:~::~:;·~~:i::D1m1::ttN[\f:i;"rt1
inPut-file-sPecification[qualifiersJ
outPut-file-sPecification[qualifiersJ

/PROCESS=

r~g::J
~NDEX ~

Indicates the type of sort to be performed. /PROCESS=RECORD is the
default.

Running SORT in Interactive and Batch Mode

2-17

/KEY=
This qualifier must be specified unless defined in a specification file.· It defines a sorting key, and may appear several times in a single command string
in order to specify several sort keys (up to 10).

NOTE:
ThE~ /KEY subqualifiers group must be enclosed in parentheses.

( (NUMBER=n]
n specifies the precedence of the sort key being defined, where 1 is the
primary sort key, 2 is the secondary sort key, and so on. If this option is
not specified on the first /KEY qualifier, NUMBER=! is assumed. If
this option is not specified on subsequent /KEY qualifiers, the default
NUMBER value is the NUMBER value of the previous key plus 1.
Legal values are 1 - 10.

,POSITION=n
n specifies the position of the key within each record, where the first
character of the record is 1. This subqualifier input must be specified.

,SIZE=n
n specifies the length of the sort key in either characters, bytes, or
digits, depending on the key field data type. This subqualifier input
must be specified. If the sort key data type is CHARACTER, key size
must be less than or equal to 255 characters. If the data type is binary,
key size must be 1, 2, or 4 bytes. If the data type is any of the decimal
types, key size must be less than or equal to 31 digits. The total of all
key field sizes must be less than or equal to 255 bytes. See Section 2.6
for additional key size information.

,CHARACTER
,BINARY
,ZONED
,DECIMAL
,PACKED_DECIMAL
ThiB subqualifier indicates the type of data appearing in the sort key
field. See Section 2.6 for data type descriptions. CHARACTER is the
default.

2-18

Running SORT in Interactive and Batch Mode

f,LEADING_SIGN
LTRAILING_SIGNJ

This subqualifier indicates whether the sign of a decimal data type key
appears at the beginning or end of the key. If the key data type is
DECIMAL and this option is not specifed, TRAILING_SIGN is the
default. See Section 2.6 for key descriptions.

OVERPUNCHED_SIGNJ
[,SEPARATE_SIGN
This subqualifier indicates whether the sign of a decimal data type key
is superimposed on the decimal value or is separate from the decimal
value. If the key data type is DECIMAL and this option is not specified, OVERPUNCHED_SIGN is the default. See Section 2.6 for key
descriptions.

f,ASCENDING
6DESCENDING

Indicates whether the key is to be sorted into ascending or descending
order. ASCENDING is the default value.

[ /WORK_FILES=n]
n specifies the number of temporary work files to be used during the sort.
Values of 0, or from 2 to 10 may be used. Default value is 2. 0 specifies no work
files because data will fit in real memory. See Section 2. 7 for additional
information.

[ /SPECIFICATION[=file-specification

Specifies the name of a file which contains SORT specification statements. If
this qualifier is not specified, a standard sort is performed. See Section 2.5
and Appendix D for additional information. SYS$INPUT is the default value.

[ /RSXll]
Indicates that SORT-11 (/RSXll) is to be invoked. The SORT-11 command
format and switches are not described in this manual. Refer to the PDP-11
SORT Reference Manual when using the /RSXll qualifier. VAX-11 SORT is
the default value.

Running SORT in Interactive and Batch Mode

2-19

NOTE:
Only the minimal unique abbreviated form of qualifier and
parameter inputs are required, but all four character abbreviations are accepted (for example, enter SPE= for
spgc1FICATION=SYS$INPUT).
An abbreviated example is:
$

SORT/KEY=CNUM=1,POS=12tSIZE=2,QECI)/SPE=
inPut-file-sPecification[qualifiersl
outPut-file-sPecification[qualifiersl

The actual example including defaults is:
$SORT/PROCESS=RECORD/KEY=< NUMBER= 1 ,posITION=12tSIZE=2,QECIMALtTRAILING_SIGN,OVERPUNCHED_SIGN,ASCENDING)/WORK_FILES=2/SPECIFICATION=SYS$INPUT inPut-file-sPecification[qualifiersJ
outPut-file-sPecification[c:iualifiersl (Bf[)

2.4.2 Input-Fiie-Specification Quallflers
Defines input file attributes.
Format:
$

SORTCc:iualifiersJ inPut-file-sPecificationUWSiO.!~iMAJhi:::~}R!t;:O:iJJ~:(l:fil§:jJ:j(g:E::#,fo@mt:::tJL§kH~tUG)i#A\i}]J

outPut-file-sPecification[c:iualifiers]

NOTE:
If the input file name does not contain a file type, the default

file type becomes DAT.
If only one FORMAT subqualifier is specified, the parentheses

() can be omitted.
[ /FORMAT=(RECORD_SIZE=n J
This input should be used only to override the record size input normally ·
retrieved from RMS. Omitting RECORD_SIZE indicates that the file record
format is to be obtained from the file header or label. n specifies the longest
record length (LRL) in bytes. The LRL input is optional, but should be
specified if the input file is not on disk or is inaccurate. The longest record
length allowed is 16,383 bytes (not including control bytes). For additional
information on determining the LRL, refer to the $FAB MRS parameter in
the VAX-11 Record Management Services Reference Manual. Note, stream
format is not supported because VAX-11 RMS does not support it.

2-20

Running SORT in Interactive and Batch Mode

~FILE_SIZE=n)]

This input should only be used to supply the file size normally provided by RMS when the input file is not on disk. This input is used to
determine the size of the work files based on input file size. n specifies
the input file size in blocks. Default is 1000 if file size cannot be
obtained from RMS and is not specified by the user. Maximum file
size is 4,294,967 ,295 blocks.

2.4.3 Output-File-Specification Quallflers
Defines output file attributes.
Format:
$ SORT[qualifiersJ

inPut-file-sPecification[qualifiersJ
outPut-file-sPecification-

!~lllllillllWPdDllli~\~l~'l'~'~~~IBllMlllllllllllll
NOTE:
If the output file name does not contain a file type, the output

file type becomes the same as the input file type.
If only one FORMAT option is specified, parentheses() may be

omitted.
[/FORMAT=]

(FIXED=n

(VARIABLE=n
(CONTROLLED=n

Indicates the output file record format. n specifies the longest record
length (LRL) of the output records in bytes, and is optional. The
longest record length allowed is 16,383 bytes (less any control bytes).
Default is input file record format if record or tag sort, and FIXED if
index or address sort. For additional information on determining the
LRL, refer to the $FAB MRS parameter in the VAX-11 Record Management Services Reference Manual.

[,SIZE=n]
This input applies to CONTROLLED records only. That is, variable
with fixed-length controlled (VFC) records. n specifies the size in bytes
of the fixed portion of controlled records. Maximum fixed control area
size is 255 bytes. If CONTROLLED is specified, and SIZE is not,
default is two bytes.

Running SORT in Interactive and Batch Mode

2-21

GBLOCILSIZE=n

)]

This input applies to magnetic tape files only. n specifies the block
length in bytes of the output file. Default value is the block size of the
input tape file, or that which was established at tape mounting time.
Block length must be in the range of 18 to 65,535 bytes.
NOTE:

To ensure for correct data interchange with other DIGITAL
systems, you should specify a block size less than or equal to
512 bytes. To ensure compatibilty with most non-DIGITAL
systems, the block size should be less than or equal to 2048
bytes.

/SEQUENTIAL
/RELATIVE
[ /INDEXED-SEQUENTIAL

Indicates the organization of the output file. If /INDEXED_SEQUENTIAL
is specified, the output file must already exist and must be empty; therefore,
/OVERLAY must be specified. Default is the input file organization if a
record or tag sort (/PROCESS= RECORD or /PROCESS=TAG) is performed.
Otherwise, /SEQUENTIAL is default.
[ /ALLOCATION=n]
n specifies the number of 512-byte blocks of disk space to be allocated for the
output file. The default value is whatever the output requires based on the
number of records sorted. Blocks allocated must be in the range of 1 to
4,294,967, 295.
[/CONTIGUOUS]
Indicates contiguous allocation of blocks for output file. Default is /NOCONTIGUOUS. This qualifier is invalid if /ALLOCATION is not specified, or if
/ALLOCATION value is insufficient for total output and the file must be
extended.
[/OVERLAY]
/OVERLAY indicates that an existing file which has the same name as the
output file should be overwritten with the SORT output. /OVERLAY requires
that the existing file must be empty. Default is /NOOVERLAY.
[ /BUCKET._SIZE=n]
n specifies the RMS bucket size (that is, the number of 512-byte blocks per
bucket) for the output file. If the output file has the same organization as the
input file, the default value is the same as input file bucket size. The maximum number of blocks per bucket is 32. If the output file organization is
different from the input file organization, the default value is 1.

2-22

Running SORT in Interactive and Batch Mode

2.5 Specification Fiie
Use of the /SPECIFICATION qualifier in the SORT command allows SORT
to be controlled by SORT specification statements. These statements are
contained in the header record and field records of a specification file, and
provide a means for expanding the range of sorting features.
Having sort processes controlled by specification files enables dynamic program control of specification file statements, and therefore dynamic control of
subsequent sort processes using the same specification file modified. Also,
specification file libraries can be maintained for often-used sorts.
The command string for a typical standard sort using a specification . file
would look like this:
$

SORT/SPECIFICATION[=specification file]
input-file-specification output-file-specification

There are eeveral methods of entering the /SPECIFICATION qualifier. If you
allow the predetermined specification file statements to control the sort,
SORT will run automatically with no further operator prompts.
Example:
$

SORT/SPECIFICATION[=file-sPecification
of the Predetermined specification file]
inPut-file-sPecification
output-file-specification CB~

However, if you use the default specification file (that is, =SYS$INPUT, and
providing your terminal is set to be the input device), SORT will prompt you
for the specification file values.
Example:
$

SORT/SPECIFICATION
inPut-file-sPecification
OUtPUt-file-SPecification (BTI)

PLEASE ENTER SPECIFICATION FILE RECORDS.
_enter the specification file header record 1.iali.tf?~5 IBID
-enter the specification file field record values for
the 1st ~\e'/ field CBtf)

_enter the specification file
t he

l a '::i t

K e '/

f i e 1d

field

record values far

(QIB~I]

Running SORT in Interactive and Batch Mode

2-23

If you allow SORT to prompt you for the input and output files as well as the
specification file values, then the SORT command will be input in the following sequence:
SORT/SPECIFICATION ~
_file: input-file-specification ~
_OUTPUT: output-file-specification ~
PLEASE ENTER SPECIFICATION FILE RECORDS.
-enter the specification file header record 1.1alues ~
-enter the specification file field record values for
the 1st ~\e)' field~
$

-enter the specification file field record values for
the last ~\e)' field@~

2.5.1 Specification Fiie Records
The specification file records can have either of two formats; fixed position
field format (SORT-11), and logical position field format (VAX--11 SORT).
NOTE:
Since omit/include and alternate collating sequences are not
supported, ALTSEQ records and record type records are invalid and cause errors. Only header and field specification records are processed.
Fixed Position Fleld Format (SORT-11)

In order to allow ease of conversion from SORT-11 V2 use to VAX-11 SORT,
the existing fixed-position-fields format of SORT-11 is accepted.
Free Fleld Format (VAX-11 SORT)

To allow some flexibility for new users, fields may be separated by commas,
and records may be variable length up to 132 characters. Blanks are ignored
unless they are embedded within a field, such as 1 00, in which case an error is
generated. Continuation lines are supported as in DCL. The individual fields,
their length, meaning and order are identical to SORT-11.
Comments in this format are placed at the end of the line by placing an
exclamation point ( ! ) immediately before the comment. The format for a
VAX-11 SORT header record would look like:
Pase nu1r1bertline nu1r1ber,H,sort t)'Pettotal ~\e)' field size,
sort ins order ,col la.tins sec:iuence ,output ~\e)' trecord lensth
! c o lrlfrl e n t

2-24

Running SORT in Interactive and Batch Mode

A specific example would be:
1 t1 tHt t10tAt t}-{t132
!header for
ordert KeY field size 10.

record sort ascend ins -

2.5.2 Specification Fiie Record Formats
Two record types and two record formats exist for specification files. The two
record types are header records and field specification records. Each record
type may contain either fixed position fields to support SORT-11 compatible
files, or free fields for VAX-11 SORT.
A DIGITAL SORT
specification form is available for use when setting up fixed position fields (see
Figure 2-6).
2.5.2.1 For SORT-11 Type Flies (fixed position flelds) -

Figure 2-6: SORT Specification Form

SORT SPECIFICATIONS
D11e _ __

Paye

HEADER SPECIFICATION
Programmer-------·----

!
>-

SORTR
SORTT
SORTA
SORTI

P•og•am

ldent1ficat1011

_ -· _____ --·· _.

---.--------------

SORTR,T

Mode of Processin

Lone

o l [J I]
!~ ~6. .1]_ 78 7! l!O

ITT

To111
Leng1h of
Key Fields

NOT USED

Comments
(Program 1ft'cntif1ca11onl

Comments

FIELD SPECIFICATION

.-------...---..---------- ·---

.----.--...........

----·--------~

Forced

Lme

~za

Field Location

~~i

~ ! §1--Fr-om·~-To----1~ ! 8

I-----,
Field Nilme

Comments

Running SORT in Interactive and Batch Mode

2-25

The format of each type of SORT-11 specification record is fixed. The SORT
specification form is based on card columns, as shown in Figure 2-6. The
following entries are common to both types of specification file lines.
Column

Notes

Entry

1-2

Page number

Required only when different types of records are to be
described. A separate page, numbered in ascending sequence,
should be used for each record type and its corresponding Field
Spe~ifications. Only the first page has a header specification.

3-5

Line number

Specifies line sequence. If column 5 is blank, 0 is assumed.
Thus a digit entry in this column can be used to identify later
line insertions.

Specification Type

H for Header, or F for Field

In the following material, unless otherwise stated, these criteria apply:
• Numeric data is decimal.
• Either leading zeroes or leading blanks are acceptable in right-justified
entries.
• All field position definition records begin at column 1.
Table 2-3 summarizes all fixed position SORT specification entries.
Header Record:

The first record in a specification file must be the header. The header tells the
SORT program what kind of sorting process to use, key field size, sorting
order, and output record size.
Format:

2-26

Field
Position

Function

Legal Values

1-2

Page number

Any number or blanks

3-5

Line number

Any number or blanks

Header record ID

7-12

Sorting process

SORT R,l,A, T or blanks

13-17

Total key field size

Any number or blanks

Sorting order

A, D, or blank

19-25

(not used)

Blanks (anything-ignored)

(not used)

Blank (anything-ignored)

(not used)

Blank (anything-ignored)

(not used)

Anything-ignored

29-32

Output record length

Any number or blanks

33-132

(not used by SORT, may be
used for comments)

Anything-ignored

Running SORT in Interactive and Batch Mode

Notes and Comments on Header Specification Entries:
Columns

Explanations and Legal Entries

7-12

Type of SORT (must be left-justified)
Legal values: SORTR or blanks - Record sort
SORTT - Tag sort
SORTA - Address sort
SORT! - Index sort

13-17

Total of all key field sizes
Legal values: 1-255
Must be equal to the total size in bytes of the largest record
key on the file and right-justified.

Normal sort order sequence
Legal values: A or blank - ascending
D - descending
This field may be qualified by N or 0 entered in column 7 of the field
specification.

29-32

Output Record Length (for SORTR and SORTT only)
Legal values: A decimal number (right-justified) equal to the number of
bytes in the largest output record.
To determine this number, add the sizes of the key fields in the field specifications for the largest record in the file. If neither SORTT nor SORTR are to
be run at this time, an entry in this field is not needed.

Field Specification Records:

The field specification records follow the header record and specify key fields
(up to ten).
NOTE:
Data fields are not supported since each entire record in a file is
written to the output file for SORTR and SORTT. For SORTA
or SORTI, output files contain only pointers and possibly some
restricted-format key data.
Format:
:Field
Position

Function

Legal Values

1-5

Page/Line number

See Header Specification

Field record ID

Key field order

Nor 0

Key field type

B,C,D,I,J,K,P,Z

9-12

First byte of field

Any number or blanks

13-16

Last byte of field

Any number or blanks

17-19

(not used)

Anything-ignored

20-80

(not used - available
for comments)

Anything-ignored

Running SORT in Interactive and Batch Mode

2-27

Notes and Comments on Field Specification Entries:
Columns
7

Explanations and Legal Entries
Key Field Order - specifies keys and their -sort sequence (this entry can satisfy
the column 18 entry for the Header Specification).
Legal values: N - normal sort sequence
'
0 - opposite sort sequence

Key field data type codes:
B - Binary (two's complement binary)
C - Character (8-bit ASCII coded alphanumeric characters). This is the default data type.
D - Decimal d~ta with sign trailing and overpunched.
I-

Same as D, but with the sign leading and separate, so that the first byte of
the field is a + or -.

1J - Same as I, but with the sign trailing and separate.
K - Same as D, but with the sign leading and overpunched
P - Packed-decimal format.
Z - Zoned ASCII format.
9-12

Field location (location of the first byte of a multi-byte key field).
Legal values: A decimal number (right-justified) specifying the first byte of
a key field.
Blanks can be used to specify a one-byte key field.

13-16

Field location (location of the last, or only, byte of a key field).
Legal values: A decimal number (right-justified) specifying the last byte,
or the only byte, in a key field.

Table 2-3: Fixed Position SORT Specification Summary
Header Specifications

---

Entry

Column

Explanation

Header specification

7-12

SORTR
SORTT
SORTA
SORTI

Record sort
Tag sort
Address sort
Index sort

13-17

1-255

Decimal number specifying the total length of all key
fields listed in the Field Specifications (must be the
maximum for SORTR).

A or blank,

Sort processing sequence: ascending or descending.

29-32

Decimal number
(SORTR or SORTT
only) 1-16,383.

This entry specifies the number of bytes for the largest
record.
(continued on next page)

2-28

Running SORT in Interactive and Batch Mode

Table 2-3: Fixed Position SORT Specification Summary (continued)
Field Specifications
Explanation

Entry

Column
6

Field specification

Normal - Key field sequenced as indicated in column
18 of Header Specification.

Opposite - Key field sequenced opposite to column 18
of Header Specification.

Character type data (8-bit ASCII alphanumeric data
in key field).

Zoned ASCII.

Digit - use digit value or convert to binary for FORTRAN IV numbers.

Same as D, but with sign leading and separate (that is,
the first byte of the field is a + or - ).

Same as D, but with sign trailing and separate (that is,
the last byte of the field is + or - ) .

Same as D, but with sign leading overpunched (that is,
the sign is superimposed on the first byte of the field).

Packed-decimal data type.

Binary data type - the key field is in two's complement
binary notation.

9-12

Decimal number
1-16,383

Location of the first byte of the key field.

13-16

Decimal number
1-16,383

Location of the last (or only) byte in the key field.

17-19

(not used)

All values are ignored.

20-80

Anything

Comments.

Sample Fixed Position Specification Fiie:

The following sample shows a header record and field record. Together they
specify an index sort process on a character key of four bytes starting in
position 10 of the record, and the output file is to be sorted in descending
order.
c 0 11.llllrl numbers
Header Record
Field Record

1234587
11
HSORTI
FOC00100013

33
48
HEADER I NOE)-{ SORT ALL DEFAULT
FIELD SPEC ALPHA KEY

NOTE:
This sample specification file performs the same sorting process
as the sample shown for free field position format.

Running SORT in Interactive and Batch Mode

2-29

2.5.2.2 For VAX-11 SORT (free fields, that Is flelds separated by commas): -

Free fields are formatted in the same sequence as the fixed position fields
described previously; however, instead of identifying fields with column numbers, commas are used. If you wish to enter blanks or use the default value,
. you must follow the entry with a comma.
Header records and field specification records are used in the same manner
here as they are for the fixed position field records described previously, and
the same explanations and legal entries also apply (see Table 2-3).

Header Records:
Field
Position

Function

Page number

Any number, or blank, or comma,

Line number

Any number, or blank, or comma,

Header record ID

Sorting process

SORT R,I,A, T, or blank, or comma

Total key field size

Any number, or blank, or comma,

Sorting order

A, D, or blank, or comma,

Collating sequence

Anything, or comma,

Output key

Anything, or comma,

Record length

Any number, or blank, or comma,

Comment

! anything, or blank

Legal Values

Fleld Specification Records:
Field
Position

Function

Page number

See Header

Line number

See Header

Field record ID

Key sorting order

N, 0, or blank, or comma,

Key data type

B,C,D,I,J,K,P,Z, or blank, or comma,

start position of key

Any number, or blank, or comma,

ending position of key

Any number, or blank, or comma,
Anything-ignored, or comma,

8
9

2-30

Legal Values

comment

Running SORT in Interactive and Batch Mode

! Anything-ignored

Sample Free Fleld Position Specification Fiie:

The following sample shows a header record and field record. Together they
specify an index sort process on a character key of four· bytes starting in
position 10 of the record, and the output file is to be sorted in descending
order.
Header Record t t H t SORT I t t , t , ! HEADER INDEH SORT ALL DEFAULT E}<CEPT
TYPE
t1F101C110t13tt!FIELD SPEC ALPHA KEY 4 BYTES POS 1 (I
OPPOSITE ORDER

Field Record

NOTE:
This sample specification file performs the same sorting process
as the sample shown for fixed position format .

2.6 Setting Up the Keys
When entering the SORT command qualifier /KEY, you must specify
/KEY=(subqualifiers and values) for each key field by which the records are
to be sorted. The total size of all key fields must be less than or equal to 255
bytes, and the maximum number of key fields allowed is ten.
Before entering the key subqualifiers and values into the SORT command
string, perform the following steps (1through6). See Section 2.4.1 for /KEY=
specification information. Figure 2-7 provides a flowchart for quick reference
when setting up keys.
Step 1

Each sort key is assigned a precedence number. You may choose to use either
the default system or the key numbering system. First decide what your key
fields will be and in what order you want them sorted. Make notes of their
sequence for use when assigning the precedence number to each. Note, if you
intend to enter the sort keys into the command string in the order of precedence you have chosen for your sorting operation, then the NUMBER=
subqualifier is not necessary. Instead, the default feature will automatically
assign the first key entered as key number 1 and each subsequent key the next
higher number.
K

NUMB:a~p,le:
18

.no7required

~:'3'.

Q-MK-00020-00

2nd. Key 3rd. Key

4th.Key

$ SORT/KEY=()1K_E_Y-==(-)/KEY=( )/KEY==(· )

(default)

2nd.Key

Primary .Key

4th. Key

3rd. Key

Running SORT in Interactive and Batch Mode

2-31

Figure 2-"/: Setting Up the Keys

Make sure to enter keys
Into the SORT command
string In the correct
order or .precedence If
default , key numbering
ll' used.

NO
(default

STEP 1
ASSIGN
KEY
NUMBERS

Make notes of key number
assignment, starting position,
and key field size for each key.

CHARACTER
(default)

BINARY
SPECIFY
BINARY

YES

SPECIFY
ZONED

STEP 5
SPECIFY START
POS and SIZE
for each key

STEP 8
SPECIFY
ASCEND* or
DESCEND ORDER

STEP3
SPECIFY
LEADING or
TRAILING SIGN*

TEP4
SPECIFY
OVERPUNCHED*
or SEPARATE SIGN

NOTE: Data type determines the
unit of key field size.
NOTE: Total size of all key fields
must be le11 than or equal
to 255 bytes.
*Indicates default parameter.

Enter these key specifications
for each key Into the SORT
command string.

2-32

Running SORT in Interactive and Batch Mode

F·MK-00021·00

Step 2

Before you can enter key size in Step 5, you must first determine what the
data type of each key field is. Figure 2-8 shows a summary of the data types
supported. See Appendix C for specific descriptions of the date types.
Example:
For decimal data type, specify DECIMAL in the command string. For example:$ SORT/KEY=(POS=l,SIZE=lO,DECIMAL).

Figure 2-8: Recognizing Data Types and Signed Numbers
1. CHARACTER -

2. BINARY -

8-b It ASCII coded alphanumeric characters.

(for example, 01010101)
Packed-Decimal String
POSITIVE NUMBER
Odd#
Even#
54321+ II
54321A

NEGATIVE NUMBER
Odd#
Even#

4321+ Is
04321A

54321- Is
54321D

TT._______

4321- Is
04321D

Last byte In string
.___ _ _ _ _ _ First byte In string

3.. DECIMAL -

8-blt ASCII coded decimal digits 0-9

~•ding

Numeric String

(Moat significant byte ls llgnod 12345)

SIGNED NUMBER
+or -12345

UNSIGNED NUMBER
12346

Separate Numeric Format

Overpunched Format
+ 12346 Is A2345

+12345

~ex 28 or 2~

-12346

-12346 II J2345

Tralllng Numeric String (Least significant digit la algned 64321)

SIGNED NUMBER
54321+ or -

UNSIGNED NUMBER
54321

Separate Numeric Format

Zoned Numeric Format

Overpunched Format

64321+

54321 + Is 54321

54321 + Is 5432A

54321- Is 5432g

54321-11 5432J

/Hex 28 or 2d\

54321-

~
~

Running SORT in Interactive Batch Mode

2-33

Step 3

If data type is decimal, specify the position of the sign. If data type is not

decimal, proceed to Step 5.
Examples:
1. For +12345 or -12345, specify the optional keyword LEADING_SIGN in
the command string. For example:
$ SORT/KEY=<POS=1 tSIZE=10tDECIMALt LEAD)

2. For trailing sign numbers (that is, 12345+ or 12345-), the optional key-

word TRAILING_SIGN is not required (TRAILING_SIGN is default).

Step 4

For decimal data types, specify if the sign is overpunched (superimposed) or
separate from the decimal value.
Examples:
1. For separate sign (that is, +12345 or -12345), specify the optional keyword

SEPAHATE_SIGN in the command string. For example:
$ SORT/KEY=<POS=1 tSIZE=10 tDECIMAL tLEAD tSEPA)

2. For overpunched sign (that is, 5432A or 5432J), the optional keyword
OVERPUNCHED_SIGN is not required (OVERPUNCHED_SIGN is
default).

Step 5

You must specify the starting position (first character in the key field) and the
size for each key. The first character of the record is 1.
Example: (in this example, key NUMBER= is default, data type is
CHARACTER).

$ SORT/KEY =(POS== 1,SIZE=10 )/KEY =(POS=22,SIZE=8 )/KEY =(POS=42,SIZE=6· )/KEY=-(POS=:84,SIZE=18·)

80-~:::r r i I~I ::: ~: := [ ~
1
10
Character/ ~
POBitione
10-ch.

key

79 80

........,,.,,,,48

8-ch.

7-ch.

16-ch.

key

Q..MK-OOOU-00

2-34

Running SORT in Interactive and Batch Mode

NOTE:
Key field size can represent either the number of bytes, or
digits depending on the data type. The chart below describes
which unit of key field size to use:
Data Type

Size Indicates

Character

Number of characters (bytes) must be less than or equal to
255.

Binary

Number of bytes must be either 1, 2, or 4.

Decimal

byte for decimal values in the range of -128 to 127.

bytes for decimal values in the range of --32, 768 to
32,767.

bytes for decimal values in the range of-2,147,483,648 to
2,147,483,647.

Number of digits in the string must be less than or equal to 31.

The number of bytes in the key field for character, binary, zoned numeric,
and overpunched is identical to the number of characters or digits. The size of
leading separate or trailing separate fields is equal to the number of digits
plus one. The size of packed-decimal fields is equal to (number of digits/2) + 1.
See Appendix C for additional information.

Step 6

Specify for each key, whether that key is to be sorted into ascending or
descending order. Ascending order is default.
Example:
To sort the first key in ascending order and the second key in descending
order, enter the key parameters into the command string as follows:
$

SORT/KEY=(POS=11SIZE=10)/KEY=(POS=221SIZE=81DESCENDING)

Now that you have performed Steps 1through6 to assemble specifications for
each key, you are ready to _enter these key specifications into the SORT
command string.

2. 7 Setting Up the Work Files
SORT automatically assigns two work files to your SYS$DISK device if you
choose to use the default. The size of these two work files (SORTWORKO and
SORTWORKl) is determined by SORT from the size of your input file.
(Note; if no assignment is done, work files are created on SYS$DISK).

Running SORT in Interactive Batch Mode

2-35

To assign your work files to a device other than the deviee your directory is on,
type:
$ASSIGN (device>: SORTWORKO
$ASSIGN (device>: SORTWORKl

$ASSIGN (device>: SORTWORK8

Example: $ ASSIGN 063: SORT WORK 1
Figure 2-9 illustrates how logical names are assigned to physical devices.

Figure 2-9: Specifying W~rk Files.
WORK FILE #

Logical Name

1
2
3
4
5
6

SORTWORKO
SORTWORK1
SORTWORK2
SORTWORK3
SORTWORK4
SORTWORK5
SORTWORK6
SORTWORK7
SORTWORK8
SORTWORK9

7
8
9
10

A specific physical device code is assigned to a
specific logical name using the ASSIGN, or DEFINE commands.

Physical Device C o d e s - - - - - - - - - - - DB: RP04, RP05, RP06 Disk
RK06 Disk

OM:

Example:

ASSIGN

ORKO
Unit 0

Default is AO
Controller A

- - - - - - RP06 Disk

2-36

}

Running SORT in Interactive and Batch Mode

Chapter 3
Calllng SORT from User Programs

You can use SORT as a set of callable subroutines from your programming
language. There are two functional interfaces to choose from; the file 1/0
interface and the record 1/0 interface. Both 1/0 interfaces share the same set
of six subroutines, and the same calls are used from all languages.
This SORT subroutine package consists of six external function calls. Each
call causes a phase of the SORT program to be performed, and returns a
status (32-bit) value indicating either success or the failure type of the phase.
Calls and associated parameters conform to the VAX-11 standard calling
interface. The calls are:

Subroutine Name

Function

1. SOR$INIT_SORT

Initialize scratch files, work area, sorting parameters

2. SOR$PASS_FILES

Pass a file name to SORT

3. SOR$RELEASE-REC

Pass a record to SORT

4. SOR$SORT_MERGE

Initiate sorting and intermediate merging of records

5. SOR$RETURN-REC

Initiate final merge pass and receive output record from
SORT

6. SOR$END_SORT

Allow clean up of files and work area to complete the
sort operation

3.1 Fiie 1/0 Interface
The file 1/0 interface enables you to specify an input file and an output file to
SORT. SORT then reads the data from the input file and sorts it into the
output file.

3-1

For the file VO interface, use the following four calls in the order listed:
Function

Call
1. SOR$PASS__FILES

Pass file specifications

2. SOR$INIT__SQRT

Initialize work areas

3. SOR$SORT._MERGE

Sort records

4. SOR$END__SORT

Clean up work areas

3.2 Record 1/0 Interface
The record 1/0 interface enables you to pass individual data records to SORT.
SORT orders them, then returns each record in correct order, individually.
For the record VO interface, use the following five calls in the order listed:
Function

Call
1. SOR$1NIT__SQRT

Intialize work areas

2. SOR$RELEASE_REC

Pass an input record

3. SOR$SORT_MERGE

Sort records

4. SOR$RETURN_REC

Receive a sorted output record

5. SOR$END__SORT

Clean up work areas

NOTE:
Calls 2 and 4 are each repeated for as many times as there are records to be
sorted.

3.3 Programming Considerations
Any program can use either SORT subroutine package interface, providing
the languagEi used produces VAX-11 native mode code and supports the following features.
• 32-bit integers
• Longword addresses
• Call by string descriptors
• Call by reference
• Either CALLS or CALLG (that is, VAX/VMS standard calling sequence)
• External function calls (each SORT subroutine returns a 32-bit status code)

3-2

Calling SORT from User Programs

Additional information regarding the VAX/VMS calling standards can be
found in Appendix C of the VAX-11 Common Run-Time Procedure Library
Reference Manual. SORT follows the Mod\Ilar Procedure Standards and uses
the common Run-Time Library routines to allocate memory and event flags.
However, SORT is not re-entrant.
The SORT subroutines are a part of the standard VMS library, therefore to
use the package a user only has to code the appropriate calls into his program,
compile or assemble and link. During the linking process the appropriate
SORT routines will automatically be linked with the user's program.
Figure 3-1 summarizes the callable subroutine set.

Figure 3-1: Subroutine Set Summary
NOTE:.
Use the subroutine calls In the order shown.

Call

Function

1. SOR$PASS_FILES

Open the Input file and create the output file.

2. SOR$1NIT_SORT

Set up the key comparison buffer and validate key
Information.
Get memory for sorting Initial phase, Input and output
buffers, and set up to read Input.
Create work files and lnltlallze the sort.

3. SOR$RELEASLREC

Get record from user and bulld key.

4. SOR$SORT_MERGE

Insert record by key into sort tree.
If sort tree Is full, continue; If not, get another record.
Output records to work files as a number of strings of
sorted records.
Output and Input until no more records and all records
are output.
Read In strings from work file and merge them until
there are ten or less left in work files.

5. SOR$RETURN_REC

Set up to output records to user.
Do final merge pass to output records (not work files)
to user.

6. SOR$ENO_SORT

Return memory, close output and Input files, and delete work files.

Calling SORT from User Programs

3-3

3.3.1 Key Comparisons
Both Interfaces

When using either interface, you have the choice of allowing SORT to do key
comparison to determine the correct order of any two records, or of writing a
routine of your own that SORT can call to do the key comparisons.
The advantage of writing your own routine is that you may know a great deal
more about the nature of the key data and therefore write a routine specifically tailored to that particular data. Because SORT does not know anything
about the key data in advance of receiving it, SORT's key comparison routine
must be general in order to handle all types of data. A routine tailored to a
particular data type or set can therefore be much·more efficient, both in space
and performance.

If you want to use the SORT key comparison routine, you must provide the
key definitions in the SOR$INIT_SORT call. Or, if you want to use your key
comparison routine with SORT, you must pass the address of your routine's
entry point (with parameters) to SORT in the SOR$INIT_SORT call. See
Section 3.4.3, Definitions, for details. Users can write a program that uses any
key data type.
For debugging purposes, it should be noted that the key comparison routine
may not necessarily be called each time a call is made to SORT. This situation can occur with the following calls:
SOR$RELEASE-REC
SOR$SORT_MERGE
SOR$RETURN-REC
Record 1/0 Interface Only

For record 1/0 interface, you must set up the key data area before passing the
record to SORT if SORT is to do the comparisons.
The key field must be set up with each key physically next to the one before it,
in order of precedence from left to right.
For example:
If the key definitions looked like this:
Key 1 - Character, Ascending, Pos 1, Size 4.
Key 2 - Binary, Descending, Pos 15, Size 2.
Key 3 - Packed, Ascending, Pos 30, Size 4.

Then, the key area in the user's program should look like this:
,A

I 0.aj-3,6
c

3\4 D,

keyl k/y2 key3

SORT will handle ascending/descending considerations as long as SORT is
doing the key comparisons. The user does not have to modify the key data in
any way.

3-4

Calling SORT from User Programs

In addition the entire key area must physically preceed and be adjacent to the
record. For example:
(KEYAREA) ABCD45634D
(RECORD) ABCDEFGHIJKLMN45 ....... .

When passing the record to SORT, the record descriptor must describe the
entire string including the key area. Therefore, the length of the string is (total
key length plus record length) and the address is the address of the first byte
of the key area.
For record 1/0 the only valid key types are 1, 2, and 4; character, binary, and
packed decimal. However, the instruction set provides a set of decimal instructions that allow conversion from all of the other decimal formats to
packed. Therefore, when you build the key area from your record data you can
convert the other decimal types to packed, and by doing so, sort on any of the
nine valid key data types that the file 1/0 interface accepts.
When SORT returns the record it will strip off the key data. The length
returned will be the length of the record alone and the first byte of the output
buffer will contain the first byte of the record, not the key.
If you are passing the address of your own key comparison routine to SORT

and you do not wish to set up the key field :preceding the record, you may
specify a 0 value as the total key size in the call to SOR$INIT_SORT. You
then pass just the record to SORT. When SORT calls your key comparison
routine the addresses of the two keys will be the ,iddresses of the first byte of
each record.

3.4 Subroutines (Parameters, Definitions, and Valld Returns)
Each call requires several user supplied parameters. Parameters, parameter
definitions, and valid returns are provided in the following paragraphs for
each call. Both symbolic and hexadecimal values are provided for the returned messages as an aid when debugging.
All user program value parameters must be passed to SORT using "call by
reference" (that is, the address of the value in the user's data area is passed to
the SORT routine, not the value itself).
All file specifications and records are passed to SORT using string descriptors.
A descriptor is a 2-longword structure of format.
For example:
flags word

]

length. of string

address of string

= 2 words = 1 longword
= 1 longword

The address of the descriptor is passed to SORT.
To omit an optional parameter, either leave it null or pass a 0 address in the
argument list; do not pass the address of a data item with a 0 value. In general
the meaning of a parameter and its legal values are identical to the equivalent
parameter in the command line to the utility.

Calling SORT from User Programs

3-5

3.4.1 SOR$1NIT_SORT

Function: Initialize scratch files, work area, and sorting parameters.

Parameters: Each of the following parameters is numbered to match its definition which follows.
1. Key buffer address 0

2. Longest record length (LRL) •
3. File size•
4. Number of work files
5. Sort type e
6. Total key size •

7. Comparison routine address•

Notes:

o Mandatory for the file 1/0 interface and for the record 1/0 interface
only if SORT is to do the key compares.
8 Mandatory for the record 1/0 interface.

e Needed for the record 1/0 interface and input from unit record or
magnetic tape devices in order for SORT to be efficient, but is not
required.
8 Valid only for the file 1/0 interface.

e Mandatory only if parameter 1 is not present and the user program
is to do the key compares.

Definitions:

1. Set up the key buffer in your user data area. The key buffer describes the
definition of the keys to be sorted on, and has the following format:

key type
key order
start position
length

one word = 1-9 for file 1/0 interface, and 1, 2, or 4
for record 1/0 interface
one word = 0 or 1
one word = 1 to (max record size)
one word = 1-255 (depends on key type)

Up to ten of these blocks can be specified in the order of key precedence.

3-6

Calling SORT from User Programs

The key buffer must be preceded by a word specifying the number of keys
specified in the following blocks. For example:

Key 1

Key 2

number of keys

1
0

10
40

key type (character)
key order (ascending)
start position in record
length of key

4
0
60
10

key type (packed-decimal)
key order (ascending)
start position in record
length of key in number of digits

Key Types:

1 =Character
2 =Binary
·a= Zoned
4 = Packed-decimal
5 =not used

6 = Decimal leading overpunched
7 = Decimal leading separate
8 = Decimal trailing overpunched
9 = Decimal trailing separate

Key Order:

0 = Ascending

1 = Descending

When passing the address of the key buffer, pass the address of the word
with the number of keys.
2. Longest record length (LRL) is a decimal number (one word in length)
indicating the longest record length in bytes not including key size.
3. File size (one longword in length) is the value for the input file size in
blocks.
4. Number of work files (one byte in length) is the value of 2 - 10 or 0.
5. Sort type (one byte in length) is the value of 1 - 4 as listed:
1 = Record sort
2 =Tag sort

3 = Index sort
4 = Address sort

6. Total key size (one byte in length) is the value of 1 - 255.
7. Address of the user generated key comparison routine. You have the option of performing your own key comparisons, and not supplying a key
definition to SORT. SORT calls your routine at the specified address, and
with the following parameters:
1) address of key 1
2) address of key 2
SORT expects the following return value:
-1 if key 1 is less than key 2.

0 if key 1 is equal to key 2.
1 if key 1 is greater than key 2.

NOTE
Keys must not be modified in any way.

Calling SORT from User Programs

3-7

Valid Returns:
Symbolic

Hex Value

Meanin1

SOR$-SORT__ON

1C802C

A sort is already in progress or this call is in
the wrong sequence.

SOR$_MISS_.KEY

1C8004

No key definition specified.

SOR$-13AD_TYPE

1C806C

An invalid sort process was specified.

SOR$_BAD-1.JRL

1C8084

An invalid LRL was specified.

SOR$_LRLMISS

1C8074

No LRL was specified and is required.

SOR$_BAD_FILE

1C808C

An invalid file size.

SOR$_WORK_DEV

1C800C

Work file deviee not random access device or
not local node.

SOR$_VM_FAIL

1C801C

SORT failed to get needed virtual memory.

SOR$_WS_FAIL

1C8024

SORT failed to get needed working set size.

SOR$_NUM_.KEY

1C803C

Invalid number of keys specified (must be
1-10).

SOR$_KEY_LEN

1C80AC

Invalid key length specified.

SS$_NORMAL
All RMS error codes

3-8

Calling SORT from User Programs

Success
See Chapter 4.

3.4.2 SOR$PASS_FILES
Pass a file specification to SORT.

Function:

Parameters: Each of the following parameters is numbered to match its definition which follows.

]·

1. Input file descriptor

2. Output file descriptor
3. Output file organizaton

4. Output file record format
5. Output file bucket size

•

6. Output file block size

7. Output file maximum record size
8. Output file allocation
9. Output file file options

Notes:

All output file parameters are specified as for VAX-11 RMS.
O These parameters are mandatory.

8 ;These parameters are optional.
(

Definitions:

1. Input file descriptor is the string descriptor for the string in ASCII of the
input file specification.

2. Output file descriptor is the string descriptor for the string in ASCII of the
output file specification.
3. Value of output file organization (one byte in length):
FAB$C_SEQ
FAB$C-REL
FAB$C-1DX
4. Value of record format for output (one byte in length):
FAB$C_FIX
FAB$C_VAR
FAB$C_VFC

Calling SORT from User Programs

3-9

5. Value for bucket size (one byte in length) is 1 - 32.
6. Value for block size (one word in length) is 18 - 32,767.
7. Value for maximum record size (one word in length) is 1 - 16,383.
8. Value for output file allocation (one longword in length) is 1 to the maximum RMS file size.
9. Value for output file file options (one longword in length) is: see the $FAB
FOP parameters in the VAX-11 Record Management Services Reference
Manual.
Valld Returns:
Symbolic
SS$_NORMAL

Success

SOR$_SORT_ON

1C802C

A sort is already in progress or this call is in
the wrong sequence.

SOR$_VAR__FIX

1C8064

Cannot change variable records to fixed
records.

SOR$-1NCONSIS

1C805C

Inconsistent data for file.

SOR$_0PENIN

1C109C

Cannot open input file.

SOR$_0PENOUT

1ClOA4

Cannot open output file.

All RMS error codes

3-10

Meaning

Hex Value

Calling SORT from User Programs

See Chapter 4.

3.4.3 SOR$RELEASE_REC
Function:

Pass a record to SORT.

Parameters: Each of the following parameters is numbered to match its definition which follows ..

1. Record descriptor
Notes:

Parameter 1 is mandatory.

Definitions:

1. Record descriptor is the address of the descriptor for the key and record

being input to SORT. The length of the record must include the total key
length plus the total record length. Also, the key field must physically
immediately precede and adjoin the record, and the descriptor must point
to the beginning of the key.

Valld Returns:

Symbolic
SS$_NORMAL

Meaning

Hex Value
1

Success

SOR$_SQRT_ON

1C802C

A sort is already in progress or this call is in
the wrong sequence.

SOR$__BAD_LRL

1C8084

Record length is longer than LRL specified.

SOR$__BAD_ADR

1C8094

Invalid descriptor address passed.

SOR$_KEY_LEN

1C80AC

Invalid key length specified.

SOR$_EXTEND

1C80A4

Failed to extend work file.

SOR$_MAP

1C809C

Internal sort map error.

SOR$_NO_WRK

1C8014

Cannot do sort in memory, need work files.

Calling SORT from User Programs

3-11

3.4.4 SOA$SOAT_MEAGE
Function: Initiate sorting and intermediate merging of records.
Parameter1: None.

Valld Returns:
Symbolic

SS$_NORMAL

3-12

Meaning

Hex Value

Success

SOR$_SOR'f'_QN

1C802C

A sort is already in progress or this call is in
the wrong sequence.

SOR$_EXTE:ND

1C80A4

Failed to extend work file.

SOR$_NQ_WRK

1C8014

Cannot do sort in memory, need work files.

SOR$_MAP

1C809C

Internal sort map error.

SOR$_READERR

1ClOB4

Cannot read a specified input file record.

SOR$_WRIT.EERR

1ClOD4

Cannot write a specified output file record.

SOR$_BADFIELD

lClOlC

Bad data in key field.

Calling SORT from User Programs

3.4.5 SORSRETURN_REC
Function: Initiate final merge pass and receive output record from SORT.

Parameters: Each of the following parameters is numbered to match its definition which follows.

1. Record descriptor 8

2. Record size

Notes: 8 This parameter is mandatory.

Definitions:

1. Record descriptor for the output area that SORT is to place the output
record into.

2. The location (one word in length) in which SORT is to place the actual
size of the record returned.

Valld Returns:
Symbolic

Hex Value

Meaning

SOR$_MAP

1C809C

Internal sort map error.

SOR.$-EXTEND

1C80A4

Failed to extend work file.

SS$_NORMAL

Success, a record has been returned.

SS$-ENDOFFILE

870

Success, no more records to return.

Calling SORT from User Programs

3-13

3.4.6 SOR$END_SORT

Function:

Allow clean up of files and work area to complete the sort opera-

tion.

Parameters:

None

Definitions:

None

Valld Returns:
Symbolic
SS$_NORMAL
SOR$_CLEAN_UP

3-14

Meaning

Hex Value
1
1C80B4

Calling SORT from User Programs

Success
Failed to delete work files and reinitialize
work areas and data areas.

3.5 Sample MACRO Program

.TITLE TESTSUB
.IDENT x01.01
THIS IS A SAMPLE MACRO
PROGRAM WHICH CALLS
THE SORT SUBROUTINE
PACKAGE. THERE IS AN
EXAMPLE USING EACH
INTERFACE.
' DATA AREA

;

.ASCII /R010SQ.DAT/
.ASCII /TEST.TMP/
.BLKB
2
IN_FAB:
.BLKB
80
IN_RAB:
.BLKB
68
OUT_FAB:
.BLKB
80
.BLKB
68
OUT-RAB:
• LONG
10
FILEIN:
.ADDRESS FILENAMEIN
FILEOUT:
.LONG
8
+ADDRESS FILENAMEOUT
.WORD
1
KEYBUF:
.WORD
1
KEYTYPE:
.WORD
0
KEYORD:
.WORD
1
KEYPOS:
• WORD
10
KEYSIZ:
.WORD
80
INLRL:
• LONG
500,
WRKFILE:
.BYTE
ll
NUMWRK:
.BYTE
2
TAGSRT:
.BLKB
2
.BLKB
10
KEYAREA:
.BLKB
80
RECORDBUF:
.LONG
80
RECDESC:
.ADDRESS KEYAREA

FILENAMEIN:
FILENAMEOUT:

;INPUT FILENAME
;QUTPUT FI LE NAME
;RMS DATA BLOCKS

;INPUT FILE NAME DESCRIPTOR
;ouTPUT FILE NAME DESCRIPTOR
;KEY DEFINITION BUFFER

;INPUT RECORD LONGEST LENGTH
;woRK FILE SIZE
;NUMBER OF WORK FILES
;TAG SORT
;KEY BUFFER
;RECORD BUFFER
;RECORD DESCRIPTOR

FIRST THE FILE I/O INTERFACE. DO A TAG SORT ON THE FILE 'R010SQ.DAT'
INTO THE FILE 'TEST.TMP' USING ll WORK FILES. KEY IS CHARACTERt 10 BYTES
LONGt STARTING POSITION 1 •
• EXTRN

SOR$PASS-FILEStSOR$INIT_SORTtSOR$SORT_MERGEtSOR$END-SORTtSOR$RELEASE_RECtSOR$RETURN_REC

FILEID::
.ENTRY
PUSHAB
PUSHAB
CALLS
BLBC
PUSH AB
PUSHAB

FILEOUT
FILEIN
#2tSOR$PASS_FILES
ROt2$
TAGSRT
NUMWRK

;SAIJE REGISTERS
;DEFAULT ALL OUTPUT OPTIONS
;pusH FILENAME DESCRIPTOR ADDRESS
;PASS FILENAMES TO SORT
;TEST FOR ERROR
;PUSH SORT TYPE
;pusH NUMBER OF WORK FILES

Calling SORT from User Programs 3-15

CLRQ
PUSHAB
CALLS
BLBC

-<SP>
KEYBUF
•5tSOR$1NIT-SORT
ROt2$

CALLS
BLBC
CALLS
BLBC

•OtSOR$SORT-MERGE
ROt2$
•OtSOR$ENO_SORT
ROt2$

DEFAULT LRL ANO WORM FILE SIZE
PUSH KEY BUFFER ADDRESS
INITIALIZE THE SORT
TEST FOR ERROR
LET SORT DO COMPARES
START SORTING
TEST FOR ERROR
00 CLEAN UP
TEST FOR ERROR

NOW TRY THE RECORD I/O INTERFACE. RECORDS ARE 80 BYTES LONG, KEY IS
CHARACTER, 10 BYTES LONGt STARTING IN POSITION 1. WORK FILE SIZE IS
500 BLOCKS.
CALLS
BLBC

•OtOPEN-INPUT
ROt2$

PUSHAB
PUSHAB
PUSHAB
CALLS
BLBC
MOVZWL
CALLS
BLBC
MOVC3

WRKFILE
INLRL
KEYBUF
•3 tSORUNIT-SORT
ROt2$
•1000 tl~G
#OtGELRECORD
ROt2$
•lOtRECORDBUFtKEYAREA

PUSHAB
CALLS
BLBC
SOBGTR

RECDESC
•1 tSOR$RELEASE_REC
R0t2$
RS, 1$

CALLS
2$: BLBC
3$:
PUSHAB
PUSHAB
CALLS
CMPL
BEQL
BLBC
CALLS
BRB
4$: CALLS
BLBC
CALLS
MOVL
RET
6$: CLRL
RET

•OtSOR$SORT-MERGE
ROtG$

1$:

;soRT DOES COMPARES
;NO MORE RECORDS TO GIVE

INLRL
RECDESC
•2,SOR$RETURN_REC
ROtSSLENDOFFILE
4$
ROt8$
•O, PUT .. RECORD
3$
•OtSOR$ENO_SORT
ROtG$
#OtCLOSE_FILE
•1 tRO

;pusH RECORD SIZE LOCATION
;pusH RECORD DESCRIPTOR
;GET RECORD BACK
;GOTTEN ALL RECORDS
;YES
;ERROR
;puT RECORD INTO OUTPUT

;INDICATE FAILURE

.END

3-16

;oPEN USER INPUT ANO OUTPUT FILE
;TEST FOR ERROR
;DEFAULT SORT TYPE ANO WORK FILES
;pusH WORK FILE SIZE
;pusH LRL
;pusH KEY BUFFER ADDRESS
;INITIALIZE THE SORT
;TEST FOR ERROR
;sET UP LOOP INDEX
;GET RECORD FROM MY FILE
;TEST FOR ERROR
;SET UP KEY IN KEY BUFFER
;soRT DOES COMPARES
;pusH RECORD DESCRIPTOR
;GIVE RECORD TO SORT
;TEST FOR ERROR

Calling SORT from User Programs

;FINISH UP
;TEST FOR ERROR
;CLOSE UP FILES
;INDICATE SUCCESS

3.6 Sample COBOL-74NAX Program
IDENTIFICATION DIVISION.
PROGRAM-IO. TSTSORT.

*
* THIS IS A SAMPLE COBOL-74/VAX PROGRAM THAT CALLS THE NATIVE
* SORT SUBROUTINE PACKAGE USING THE RECORD I/O INTERFACE. IT

* REQUESTS A RECORD SORT USING A 5 BYTE CHARACTER KEY.

ENVIRONMENT DIVISION.
INPUT-OUTPUT SECTION.
FILE-CONTROL.
SELECT FILE-IN
ASSIGN TO 11 SY 11 •
SELECT FILE-OUT
ASSIGN TO "SY".
DATA DIVISION.

* ASSIGN FILE DEVICES
* ANO NAMES ANO DEFINE INPUT AND OUTPUT RECORD AREAS.
*
FILE
SECTION.
FD
01

FD
01

FILE-IN
VALUE OF ID IS 11 SORTIN.DAT 11
LABEL RECORDS ARE STANDARD.
IN-REC.
05 IN-1
PIC X<9> •
05 IN-2
PIC X<5>.
05 IN-3
PIC X<S>.
FILE-OUT
VALUE OF ID IS 11 SORTOU.DAT 11
LABEL RECORDS ARE STANDARD.
OUT-REC
PIC X<20>.

** SET UP DATA FOR SORT SUBROUTINE PARAMETERS.
*WORKING-STORAGE SECTION.
77
77

ENO-OF-FILE-SW
88 END-OF-FILE
SHOW-STAT

PIC X

VALUE 11 0 11 •
VALUE 1".
PIC 9(9) •
II

** LONGEST RECORD LENGTH, WORK FILE SIZE
* AND RETURN STATUS VALUES.
*77 LRL
PIC 99
VALUE 20 COMP.
77
01

FILE-SIZ
SORT-STATUS
88 SS-NORMAL
88 SS-ENDOFFILE

PIC 9(8) VALUE 1 COMP.
PIC S9(8) COMP VALUE O.
VALUE 1.
VALUE 2160.

** KEY BUFFER INDICATING ONE 5 BYTE CHARACTER KEY STARTING IN
* POSITION 10 OF EACH RECORD, ASCENDING ORDER.
*01 KEY-BUFFER.
05
05
05
05
05

KEY-NUMBER
KEY-TYPE
KEY-ORDER
KEY-START
KEY-LENGTH

PIC 9(4)
PIC 9(4)
PIC 9(4)
PIC 9(4)
PIC 9(4)

VALUE 1 COMP.
COMP
VALUE 1 •
IJALUE o.
COMP
COMP
VALUE 10.
COMP
VALUE 5.

** AREA FOR KEY AND RECORD.
*
Calling SORT from User Programs

3-17

WK-REC-ALL.
05 WK-KEYl
PIC X<5>.
05 WK-REC.
10 WK-1
PIC X<S>.
10 WK-2
PIC X<S>.
10 WK-3
PIC X<G>.
PROCEDURE DIVISION.
MAIN-LOGIC•

** OPEN THE INPUT AND OUTPUT FILES. THEN INITIALIZE THE SORT
*SPECIFYING THE KEY DEFINITION, THE LRL AND WORK FILE SIZE.

OPEN INPUT FILE-IN
OUTPUT FILE-OUT.
CALL "SOR$INIT_SORT" USING KEY-BUFFER LRL FILE-SIZ
GIVING SORT-STATUS.
IF NOT SS-NORMAL
MOVE SORT-STATUS TO SHOW-STAT
DISPLAY "FAILURE DURING SOR$INIT, STATUS WAS "SHOW-STAT
PERFORM ABORT-JOB.

** READ RECORDS FROM FILE
* EXTRACT THE KEY AND THEN HAND EACH TO SORT.

* PERFORM RELEASE-RECS UNTIL END-OF-FILE.
** END OF FILE CALL SORT TO FINISH SORTING RECORDS.
* CALL SOR$SORT_MERGE" GIVING SORT-STATUS.
11

IF NOT SS-NORMAL
MOVE SORT-STATUS TO SHOW-STAT
DISPLAY "FAILURE DURING SOR$MERGE, STATUS WAS " SHOW-STAT
PERFORM ABORT-JOB.
MOVE "0" TO END-OF-FILE-SW.
** REQUEST RECORDS BACK FROM SORT UNTIL ALL RECEIVED.

PERFORM RETURN-RECS UNTIL END-OF-FILE.

** CALL SORT TO CLEAN UP WORK AREAS.

CALL "SOR$END_SORT" GIVING SORT-STATUS.
IF NOT SS-NORMAL
MOVE SORT-STATUS TO SHOW-STAT
DISPLAY "FAILURE DURING SOR$END, STATUS WAS
PERFORM ABORT-JOB.

SHOW-STAT

** CLOSE FILES.

CLOSE FILE-IN
FILE-OUT.
STOP RUN.

** READ RECORDS AND BUILD KEY.

*
RELEASE-RECS.
READ FILE-IN
AT END
MOVE 1 TO END-OF-FILE-SW.
IF NOT END-OF-FILE
MOVE IN-REC TO WK-REC
MOVE IN-2 TO WK-KEYl
CALL "SOR$RELEASE_REC USING BY DESCRIPTOR WK-REC-ALL
GIVING SORT-STATUS.
11

3-18

Calling SORT from User Programs

IF NOT SS-NORMAL
MOVE SORT-STATUS TO SHOW-STAT
DISPLAY "FAILURE DURING SOR$RELEASEt STATUS WAS " SHOW-STAT
PERFORM ABORT-JOB.

** RECEIVE RECORDS AND WRITE THEM OUT.
*RETURN-RECS.
CALL "SOR$RETURN_REC" USING BY DESCRIPTOR WK-REC
BY REFERENCE LRL
GIVING SORT-STATUS.
IF SS-ENDOFFILE
MOVE "1" TO END-OF-FILE-SW.
IF NOT END-OF-FILE
MOVE SPACES TO OUT-REC
MOVE WK-REC TO OUT-REC
WRITE OUT-REC•
ABORT-JOB.
DISPLAY "ABNORMAL END OF JOB"•
CLOSE FILE-IN
FILE-OUT.
STOP RUN.

Calling SORT from User Programs

3-19

3. 7 Sample FORTRAN IV PLUS Program

c
c

PROGRAM CALLSORT

C
C
C
C

THIS IS A SAMPLE FORTRAN IV PLUS PROGRAM THAT CALLS THE
NATIVE SORT SUBROUTINE PACKAGE USING THE FILE I/O INTERFACE.
THIS PROGRAM REQUESTS AN INDEX SORT OF FILE 'R010SQ.DAT'
INTO THE FILE 'TEST.TMP'. THE KEY IS AN 80 BYTE CHARACTER
ASCENDING KEY STARTING IN POSITION ONE OF EACH RECORD.

DEFINE EXTERNAL FUNCTIONS AND DATA

c
c
c
c

C
C
C

CHARACTER*lO INPUTNAME
CHARACTER*B OUTPUTNAME
INTEGER*Z KEYBUF<5>
INTEGER*Z NUMWRK
INTEGER*Z ISRTTYP
INTEGER*4 SOR$PASS_FILES
INTEGER*4 SOR$INIT_SORT
INTEGER*4 SOR$SORT-MERGE
INTEGER*4 SOR$END-SORT
INTEGER*4 !STATUS

!INPUT FILE NAME
!OUTPUT FILE NAME
!KEY DEFINITION BUFFER
!NUMBER OF WORK FILES
!SORT PROCESS
!SORT FUNCTION NAMES

!STORAGE FOR SORT FUNCTION VALUE

INITIALIZE DATA - FIRST THE FILENAMES THEN THE KEY BUFFER FOR
ONE 80 BYTE CHARACTER KEY STARTING POSITION 1' 3 WORK FILES
AND AN INDEX SORT PROCESS
DATA INPUTNAME,OUTPUTNAME/'R010SQ.DAT' ,'TEST.TMP'/
DATA KEYBUF,NUMWRK,ISRTTYP/1 ,1 ,0,1 ,90,3,3/

CALL THE SORT EACH CALL IS A FUNCTION

PASS SORT THE FILENAMES

c
c

3-20

IBTATUS = SOR$PASS_FILES<INPUTNAME,OUTPUTNAME>
IF <.NOT. !STATUS> GOTO 10
INITIALIZE WORK AREAS AND KEYS
!STATUS= SOR$INIT_SORTCKEYBUF,,,NUMWRKtISRTTYP>
IF <.NOT. !STATUS> GOTO 10
SORT THE RECORDS
!STATUS = SOR$SORT_MERGE< >
IF <•NOT. I STATUS) GOTO 10
CLEAN UP WORK AREAS AND FILES
!STATUS= SOR$END-SORT<>
IF <.NOT. !STATUS> GOTO 10
STOP 'SORT SUCCESSFUL'
STOP 'SORT UNSUCCESSFUL'
END

Calling SORT from User Programs

Chapter 4
Error Conditions

You can encounter error conditions at three operating levels: first with the
VAXNMS DCL command interpreter, next with the SORT error messages,
and last with VAX-11 RMS messages.
SORT handles two basic types of errors; fatal and warning. Fatal errors (severity level F) cause SORT to halt processing; warning errors (severity level
W) cause a warning message to be output and allow sort processing to pro'."
ceed. Errors in both these categories are grouped into three classes:
• Errors caused by 1/0 or other system failures.
• Errors caused by misinformation passed to SORT as a parameter of a subroutine call.
• Errors caused by invalid data in a key field.
For the SORT utility, errors of all types and classes are signaled to the system; this signal causes a message to be output. Execution is either stopped or
continued based on the severity of the error. Execution can be resumed only if
the severity level is W (that is, code = 0).
In summary, only invalid data errors and a few RMS errors cause warning
error messages. System or 1/0 failures and bad subroutine parameters are
fatal. For additional information regarding error condition handling, refer to
the VAX/VMS System Services Reference Manual.

4-1

4.1 Command Interpreter Error Messages
In interactive mode, when you enter a command line incorrectly, the command interpreter issues a descriptive error message telling you what was
wrong. For example, if you specify more than one parameter for a command
that accepts a single parameter, you receive the message:
%DCL-W-MAXPARAMt maximum Parameter count exceeded

You must then retype the command line.
Other error messages may occur during execution of a command. These messages can indicate such errors as a nonexistent file or a conflict in qualifiers.
Not all messages from the system indicate errors; other messages are informative, or merely warn you of a particular condition.
The VAXNMS system messages have the general format:
%XXX-L-CODEt text

Descriptive comment.
- - - - - - - Shorthand code for the message text.
, - - - - - - - Severity Level:
S =Success
W =Warning
E =Error
F =Fatal
Mnemonic for the operating system program
issuing the message.

Example:
%SORT-W-CLOSEOUTt error
ification)

closin~

output CoutPut file-spec-

Because these messages are descriptive, you can usually understand what you
need to do differently when you issue the command again. But, if you do not,
the VAX/VMS Messages and Recovery Procedures Manual lists all the possible command interpreter error messages and describes what you can do to
correct a command interpreter error.

4.2 SORT Error Messages
The following VAX-11 SORT error messages are listed in alphabetic order.
All SORT error messages have the same format as command interpreter messages, that is:
%SORT-<seueritY leuel>-<code) tCtext).

4-2

Error Conditions

The following descriptions of error messages observe the following
conventions:
• Only the (code), (text) part of the message is shown in the following list.
• (filespec) ind~cates a
DB1:[153,101TEST.TMP;3

file

specification.

For

example:

• (number) indicates the user entered numeric value.
• LRL means the longest record length (specified in bytes).

6AD,....ADR t inl.1alid descriptor address specified.

You passed the subroutine package an address for a descriptor, and the
descriptor was invalid format. Character string descriptors in VAX consist
of two longwords. The first word of the first longword contains the character string length in bytes. The second longword contains the address of the·
string.
User Action: See Section 3.4 and check character string format.

6ADFIELDt (filesPect or field text that is inl.1alid) field invalid at (nur11ber>.

Bad data in key field or command. In this message, (number) indicates
the record number of the record containing bad data in hex.
User Action: Check key field data type and the starting positions and
lengths (See Section 2.6, Setting Up the Keys).

6AD_FILEt file size invalid.

You specified a negative file size or a zero file size. File size must be
greater than zero.
User Action: Specify a file size greater than zero.

6AD_KEYt invalid ~\e~· specification.

Either the key field size, position, data type, or order is incorrect within
the key definition. Positions start at one and cannot be greater than the
maximum record size. Size must be less than or equal to 255 for character
data, 1, 2, or 4 for binary data, and less than or equal to 31 for decimal.
User Action: See Section 2.6, Setting Up the Keys, and check the command string key specifications.

Error Conditions

4-3

BAD-LENt outPut record lensth less than 18 bYtes for Mastape.

Magnetic tape requires record lengths to be at least 18 bytes and no
greater than 4096 bytes.
User Action: See Section 2.4.3, and check your output file block size
parameters.
BAD-LRL t inPut file (filespec).
,Record size sreater than specified LRL.

In reading the input file, SORT encountered a record longer than the
specified LRL. The record will be truncated to the LRL and sorted.
User Action: ·Re-execute SORT with a larger LRL.
BAO_SPECt invalid specification file record.
FIELD: (record specification>.

An incorrect field was specified in the specification file record. (record
specification) indicates bad record contents.
User Action: See descriptions of specification file record formats (Section
2.5.2) and change field specifications.

BAD_TYPEt invalid sort Process.

You passed the subroutine package a sort type code of less than 1 or
greater than 4 if file 1/0 or not equal to 1 if record 1/0, or an invalid key
word in command /PROCESS. Legal values are 1-4 for file 1/0, nothing
for record 1/0, and RECORD, TAG, INDEX, or ADDRESS for command
/PROCESS parameter.
User Action: Specify a different sorting process.
CLEAN_LJp, failed to reinitialize 1,,iork area and files.·

SORT was unable to deallocate the extra virtual memory, deassign work
file channels, or readjust working set size. For the SORT utility, this is a
warning of little importance. For the SORT subroutine packages, this
could mean a failure to be able to recall SORT from the same program
until it has exited. This is an internal error.
User Action: Exit from the user program before re-executing SORT.
CLOSEINt error closins (filesPec) as inPut.

An error occurred closing an inprtt file. This message is usually accompanied by an RMS message indicating the reason for the failure.
User Action: Take corrective action based on the associated message.

4-4

Error Conditions

CLOSEOUT terror closins (filespec) as output.

An error occurred closing an output file. This message is usually accompanied by an RMS message indicating the reason for the failure.
User Action: Take corrective action based on the accompanying message.

EXTENDt failed to extend i...1orK file.

SORT failed to extend a user's temporary work file. Either the device is
full, or the user does not have extend privilege.
User Action: See Section 2. 7 and reassign work files to a different
device with more space, and make sure you have extend privilege on that
directory.

INCONSISt inconsistent data in file (filesPec).

If you specified /OVERLAY plus other· output file qualifiers, SORT will
verify that the information in the existing file matches the information
you provided. If it does not, this error message is reported. Unless you
specifically want a verification, /OVERLAY should be used without other
qualifiers.
User Action: Check the command string output file qualifiers (See Section
2.4.3).

IND-OVRt
fier.

indexed sequential

outPut

requires overlay quali-

You specified indexed output file organization and did not specify /OVERLAY.
User Action: You must create the indexed file first with RMS DEFINE
utility (or other). The primary key of the file should be the same as the
sort key for efficiency but is not required to be. Then you must specify
/OVERLAY in the SORT command string.

KE'1'-LENt •te}'
Cn1.1mber>.

lensth

in\.1alid.

Ke}'

tHlf11ber

(n1.u11ber)t

size

The key size is incorrect for the data type, or the total key size is greater
than 255.
User action: See Section 2.6, Setting Up the Keys, and specify correct key
field size. Size must be less than or equal to 255 for character data, 1, 2, or
4 for binary data, and less than or equal to 31 for decimal. Also, only
ascending or descending order is allowed.

Error Conditions

4-5

LRL_MISS t LRL rr1ust be specified•

If record 1/0 interface subroutine package is selected, the longest record
length (LRL) must be passed to SORT in the call.

User Action: See Section 3.4, and specify LRL.
MAP, failed to rr1aP 1,,1orR file.

This is an internal SORT failure.
User Action: Verify that the system parameter "maximum process sections" has been set up at 10. If it has, then report this failure to a specialist. Otherwise, set that system parameter to 10.
MISS_KEY t H.eY specification frlissins.

SORT did not find any key definition in either the command line or
specification file, or in the parameters to the subroutine package.
User Action: You must input at least one key definition in one of these
three areas.
NO_WRK t need 1,,1or•:. files cannot do SORT in rr1err1orY.

You specified /WORK-FILES=O indicating the data would fit in memory,
but the data was too large.
User Action: Either increase the working set quota, or allow SORT to use
two or more work files.
NUM_KEYt too man>' •:.e>'S specified.

Up to ten key definitions are allowed. Either too many have been specified, or the NUMBER value is wrong.
User Action: See Section 2.6, Setting Up the Keys, and check your command string key field specifications.
ONE_IN, onh· one inPut file allo1,,1ed.

SORT will take only one input file at a time.
User Action: You can concatenate files of the same organization and
record format using COPY, and then sort.
OPENINt error oPeninS' (filesPec> as inPut

An input file cannot be opened. This message is usually accompanied by
an RMS message indicating the reason for the failure.
User Action: Take corrective action based on the associated message.

4-6

Error Conditions

OPENOUTt error oPenins (filesPec> as output

An output file cannot be opened. This message is usually accompanied by
an RMS message indicating the reason for the failure.
User Action: Take corrective action based on the associated message.

READERRt error readins (filesPec>

An input file record specified cannot be read. This message is usually
accompanied by an RMS message indicating the reason for the failure.
User Action: Take corrective action based on the associated message.

SORT_ONt sort alread}' in Prosress.

You tried to call the SORT subroutine package with calls in the wrong
order, or to recall it before it finished running the previous sort.
User Action: Reorder the subroutine calls and then re-execute SORT.

VAR-FIXt
lenath.

cannot

chanse variable

lenSth

records

into fixed

You specified variable length input records and requested fixed length
output.
User Action:Output records must be variable or controlled in this case.

t.JM_FAILt failed to Set rec:iuired virtual 1rle1r1orY (nurriber).

SORT could not get the amount of virtual memory required for the sort.
(number) indicates the number of bytes needed.
User Action: If the SORT utility is being run, decrease the working set
quota; if either SORT subroutine package is being run, either decrease the
quota or return some memory to the system inside the user's program
before calling SORT.

WORK_DEl.Jt 1,..1orH file (filespec)
det.iice specified not rando1r1 access or not local.

Work files must be specified for random access devices that are local to the
CPU the sort is being performed on (that is, not on node in a network).
Random access devices are disk devices.
User Action: See Section 2.7, Setting Up the Work Files, and specify the
correct device.

Error Conditions

4-7

WRITEERRt error 1A1ritins (filespec>

An output file record cannot be written. This message is usually accompanied by an RMS message indicating the reason for the failure.
User Action: Take corrective action based on the associated message.
ws_FAILt failed to Set rec:iuired IAfOrKins set space (n•.lfrlber).

SORT could not set the required amount of real memory space. A minimum 75 page working set is needed. (number) indicates number of pages
available.
User Action: Increase the working set quota.

4.3 VAX-11 RMS Error Codes
Listed below, in alphabetic order, are the VAX-11 RMS completion status
codes. This list includes both symbolic and hexadecimal codes for error messages and success messages. These RMS codes are returned to your program
by the operating system.
All VAX-11 RMS error messages have the same format as command interpreter messages, that is:
'X.RMS-<se,.ieritY level>-( code) tCtext).

For additional information refer to the VAX-11 Record Management Services
Reference Manual.
Valld Returns for Error Messages:

4-8

Symbolic

Hex Value

RMS$_ACC

0001C002

File access error.

RMS$_ACT

0001825A

File activity precludes operation.

RMS$_AID

000183F4

Bad area identification number field in allocation XAB.

RMS$_ALN

000183FC

Invalid alignment boundary type in allocation XAB.

RMS$_ALQ

00018404

Incorrect allocation quantity in allocation XAB; the
value either exceeds the maximum allowed, or is equal
to zero for the extend service.

RMS$_ANI

0001840C

Records in a magnetic tape file are not ANSI D format.

RMS$_AOP

00018414

Invalid allocation option in allocation XAB.

RMS$_ATR

OOOlCOCC

Read error on file header.

RMS$_ATW

0001COD4

Write error on file header.

RMS$_BKS

0001841C

Invalid bucket size in FAB.

Error Conditions

Meaning

Symbolic

Hex Value

RMS$_BKZ

00018424

Invalid bucket size in the allocation XAB for relative
file.

RMS$_BLN

0001842C

Invalid value in block length field.

RMS$_BOF

00018198

File is already at beginning of the file (backspace operation).

RMS$_BUG_J)DI

0001843C

Invalid default directory. Internal VAX-11 RMS error;
no recovery possible - contact a software specialist.

RMS$_CCR

00018494

Cannot connect RAB (only one record stream permitted
for sequential files).

RMS$_CD A

0001COE4

Cannot deliver AST.

RMS$_CHN

OOOlCOEC

Channel assignment failure.

RMS$_COD

000184AC

Invalid type code in XAB.

RMS$_CRE

OOOlCOOA

File create error.

RMS$_CUR

000184B4

No current record; operation not immediately preceded
by a successful get or find service.

RMS$_DAC

0001C012

File deaccess error during a close service.

RMS$_DEL

00018262

Record accessed by RF A record access mode has been
deleted.

RMS$_DEV

000184C4

Bad device or inappropriate device type for operation.

RMS$_J)IR

000184CC

Error in directory name.

RMS$_J)ME

000184D4

Dynamic memory exhausted; occurs only if the related

1/0 segment in the control region is full and the file is
either a direct access process permanent file, or the user
has disallowed the use of the program region for 1/0
buffers to VAX-11 RMS.
RMS$_J)NA

000184DC

Error detected in the default file specification string.

RMS$_J)NF

0001826A

Directory not found.

RMS$_J)NR

00018272

Device not ready.

RMS$_]) PE

0001C03A

Device positioning error; applies only to magnetic tape.

RMS$_J)VI

000184F4

Invalid device identification in NAM block.

RMS$_ENT

OOOlCOlA

Error during file enter service.

RMS$_ENV

00018724

Environment error; the code necessary to support the
file organization or facility was not selected at system
generation.

RMS$_EOF

0001827A

End of file.

RMS$_ESA

000184FC

Invalid expanded string area in NAM block.

·RMS$_ESL

00018714

Invalid expanded string length in NAM block.

RMS$_ESS

00018504

Expanded string area too short.

RMS$_EXP

000182C2

File expiration date not yet reached.

RMS$_EXT

0001C022

File extend error.

RMS$_FAB

0001850C

Invalid FAB; block indentifier field incorrect.

Error Conditions

4-9

Symbolic

Hex Value

Meaning

RMS$_FAC

00018514

Operation not allowed by the value set in the file access
field of the FAB.

RMS$_FEX

00018282

File already exists.

RMS$_FLK

0001828A

File is locked and therefore not available.

RMS$_FNA

00018524

Invalid file specification string address in FAB.

RMS$_FND

0001C02A

Files-11 find function failed.

RMS$_FNF

00018292

File not found.

RMS$_FNM

0001852C

Syntax error in file name.

RMS$_FQP

0001853C

Invalid file processing options.

RMS$_FSZ

00018534

Invalid fixed control area size in FAB (equal to 1 for
print files).

RMS$_FUL

00018544

Device full; cannot create or extend file.

RMS$_IFA

0001C124

Illegal file attributes; file header corrupted.

RMS$_IFI

00018564

Invalid internal file identifier in FAB; must be zero.

RMS$_IMX

0001856C

More than one XAB of the same type is present for the
file.

RMS$_IOP

00018574

Illegal operation attempted:
1. block I/0 when not block I/0 access.
2. record I/O when block I/0 access.
3. rewind of process permanent file.
4. inappropriate device type or file organization.

RMS$_IRC

0001857C

Illegal record in sequential file; invalid count field.

RMS$_ISI

00018584

Invalid internal stream identifier in RAB.

RMS$_KBF

0001858C

Invalid key buffer address; not in access limits.

RMS$_KEY

00018594

Invalid record key for random operation to a relative
file.

RMS$_KSZ

000185A4

Key size not equal to 4 (relative file).

RMS$-1NE

000185BC

Logical name error; resulted in duplicates.

RMS$_MBC

00018734

Invalid multi-block count; must not be greater than
127.

4-10

RMS$_MKD

0001C032

Files-11 ACP could not mark file for deletion.

RMS$_MRN

000185CC

Illegal value for maximum record number.

RMS$_MRS

00018504

Illegal value for maximum record size.

RMS$_NAM

000185DC

Invalid NAM block.

RMS$_NEF

000185E4

Attempt to use the put service to a sequential file when
not positioned to end of file.

RMS$_NMF

000182CA

No more files for a search operation.

RMS$_NQD

000185F4

Node name error.

RMS$_0RG

0001860C

Illegal file organization.

R.MS$_FBF

00018614

Invalid prompt buffer address.

Error Conditions

Symbolic

Hex Value

Meaning

RMS$_pLG

0001861C

Error in file prologue; file is corrupted.

RMS$_pLV

0001872C

Prologue version unsupported.

RMS$_pRV

0001829A

Privilege violation; access denied.

RMS$_QUO

00018634

Error in quoted string.

RMS$_RAB

0001863C

Not a valid RAB; block identifier field incorrect.

RMS$_RAC

00018644

Illegal value in record access mode field of RAB.

RMS$_RAT

0001864C

Record attributes invalid in FAB.

RMS$_RBF

00018654

Invalid record address.

RMS$_RER

0001COF4

File read error.

RMS$_REX

000182A2

Record already exists; in a random access mode operation to a relative file, a record was found in the target
record cell.

RMS$_RFA

0001865C

Invalid record's file address contained in RAB.

RMS$_RFM

00018664

Illegal record format.

RMS$_RHB

0001866C

Invalid record header buffer.

RMS$_RLF

00018674

Invalid related file.

RMS$_RLK

000182AA

Record locked by another task.

RMS$_RMV

OOOlCOFC

Files-11 remove function failed.

RMS$_RNF.

000182B2

Record not found.

RMS$_RNL

000181AO

Record not locked.

RMS$_RPL

0001Cl04

Error while reading prologue.

RMS$_RSA

0001868C

Record stream active; an attempt was made to issue a
record operation request in an asynchronous environment to a record stream that has a request outstanding.

RMS$_RSL

0001873C

Resultant string length field of NAM block invalid.

RMS$_RSS

00018694

Resultant string area size field of NAM block is too
small.

RMS$_RST

0001869C

Invalid resultant string area.

RMS$_RSZ

000186A4

Illegal record size.

RMS$_RTB

000181A8

Record too large for user buffer.

RMS$_SHR

000186B4

Invalid value in the file sharing field of FAB.

RMS$_SQO

000186C4

Operation not sequential.

RMS$_SYN

000186D4

Syntax error in file specification.

RMS$_SYS

OOOlClOC

Error in system QIO directive.

RMS$_TMO

000181BO

Time-out period expired.

RMS$_TYP

000186E4

Error in file type.

RMS$_UBF

000186EC

Invalid user record area address.

RMS$_USZ

000186F4

Invalid user record area size.

RMS$_VER

000186FC

Error in version number.

Error Conditions

4-11

Symbolic

Hex Value

Meaning

RMS$_WER

0001C114

File processor write error.

RMS$_WLK

000182BA

Device is not write-locked.

RMS$_WPL

OOOlCllC

Error while writing prologue.

RMS$_WSF

0001871C

Working set full.

RMS$-XAB

0001870C

Not a valid XAB.

Valld Returns for Success Messages:

4-12

RMS$_CONTROLC

00010651

Operation completed under Control C.

RMS$_CON'I'ROLO

00010609

Operation completed under Control 0.

RMS$_CONTROL Y

00010611

Operation completed under Control Y.

RMS$_CREATED

00010619

File was created; not opened; used in conjunction with
the CIF option.

RMS$-KFF

00018031

Known file found.

RMS$_NORMAL

00010001

Operation successful (synonym for RMS$_SUC).

RMS$_0K-ALK

00018039

Record already locked.

RMS$_0K-1>EL

00018041

Deleted record accessed correctly.

RMS$_0K_RLK

00018021

Record locked but read anyway; locked set RLK bit in
ROP field.

RMS$_01L.RNF

00018049

Non-existent record accessed correctly.

RMS$____FENDING

00018009

Asynchronous operation not yet completed.

RMS$_SUC

00010001

Operation successful (synonym for RMS$_NORMAL).

RMS$_SUPERSEDE

00010631

Created file superseded an existing version.

Error Conditions

Chapter 5
Improving SORT Efficiency

Users who have special sorting requirements such as very large files, storage
media contraints, and processing time restrictions can modify SORT's behavior for optimum performance. Your ability to improve SORT's performance
depends on your understanding of SORT's operational characteristics described in this chapter.
This chapter discusses:
• How the SORT program functions in each phase of operation, and what
sequence of events occur during a sort run
• How a user can improve SORT's efficiency through the use of tuning
procedures

5.1 Functional Description
The SORT program consists of two basic parts: a control program called the
utility and a callable subroutine package (see Figure 5-1). The utility directs
the overall processing. The callable subroutine package serves as a collection
of subroutines that the utility uses during its processing. You can write your
own control program to take advantage of SORT's callable subroutines (see
Chapter 3).
There are eight phases of operation in the SORT utility. These are described
in more detail in Section 5.1.2. A sort run breaks down into three tasks.
First, SORT reads the command string and the specification file, if present,
decodes them, and then stores the qualifier values and parameters. Any errors
in the command string or specification file are reported at this point.

5-1

Figure 5-1: VAX-11 SORT Architecture, Main Functional Components
CALLABLE
SUBROUTINE PACKAGE

UTILILTY

1/191_

__,A'-'----..

MAIN LINE

RMS
FILE 1/0

COMMAND
INTERPRETER
INTERFACE

VAX-11 SORT

COMMAND

DCL
COMMAND
USER

F-MK-00024-00

5-2

Improving SORT Efficiency

Second, SORT begins the pre-sort operation. The control program calls
routines to open and read the input file and establish -the keys. Then the
SORT subroutine package is called to begin the initial sorting process. At this
point, the amount of available internal storage space becomes important to
the efficiency of the sort. If that space is not sufficient to hold all the records,
SORT builds strings of sorted records· and transfers them to work files on
temporary storage devices (disk). The SORT program normally provides for a
default of two work files. A qualifier in the command string can increase the
number of work files used.
Third, SORT rebuilds the intermediate work files into a merged file. If the
process is tag sort, another subroutine reads the records in the proper
sequence. The records are then written in the output file. If there· are no work
files to merge because main memory was sufficient to hold all the records, the
sorted records are written directly into the output file. After the last record is
written, the control program cleans up the work files and exits; SORT is then
ready to accept another job.

5.1.1 Sorting Processes
All four sorting processes can sort records of fixed or variable length, VFC, or
any valid VAX-11 RMS. Stream format is not supported. The size of the
records on a fixed-length format file is determined when the file is created.
The first word of a variable-length format record contains the size of the
record in bytes. This first word is used by the file system and is transparent to
SORT.

5.1.1.1 Record Sort - Record sort outputs all data records in a specified
sorted sequence. Each record is kept intact throughout the entire sorting
process. Since this process moves the whole record, it is relatively slow and
may require considerable· main memory or external storage work space for
large files.

5.1.1.2 Tag Sort - Tag sort produces the same kind of output file as record
sort, but it only handles record pointers and key fields. Since this process
moves a smaller amount of data than record sort, it may perform a faster sort
than record sort. The input file must be randomly re-accessed to create the
entire output file, which may be a lengthy process for large files.

Input Data Flies

A record is usually divided into several logical areas called data fields. The
data in each field may or may not be relevant to SORT. Each field may be
interpreted as a record identifier, key data, or general data related to the

Improving SORT Efficiency

5-3

logical content of the record and not relevant to the sorting process. SORT
uses record identifiers to distinguish the various types of records in a file.
SORT uses the key fields in each record to reorder an input file. Any other
data field in a record may be retained in the output file or ignored by SORT.
Figure 5-2 shows three different types of input records, each with a different
format. The record identifiers· are the letters in position 1: S means sales
record, 0 means order record, and R means restock record. In this case, the
keys chosen for sorting the sales record types are the "item number code" in
positions 2 to 7, and the "number of items sold" in positions 8to13. The "total
amount of sale" is an example of a data field not relevant to the sorting
process.
If you request a sort in ascending order on the sales records as shown in Figure
5-2, the sort is based on the item number code first and then on the number of
each item sold within that item number. In order of decreasing significance,
the keys are:
1. Item number

2. Number of items sold

Figure 5-2: Sample Record Types

RECORD
IDENTIFIER
KEY FIELD

KEY FIELD

DATA FIELD

SALES
RECORD
'--~........-~--"--~-.---''--'--.....-......_~--~......._~~-v---~--~-v~~~-----~~

NUMBER
OF ITEMS
SOLO

ITEM
NUMBER
CODE

UNIT
CODE

TOTAL
AMOUNT
OF SALE

PRICE
PER
UNIT

CUSTOMER
NO.
CODE

DATE
OF
SALE

ORDER
RECORD
'-------...----·--··-____,.__--~-·

OE SCRIPTION
OF
ITEM

-......,----____..._
ITEM
NO.
CODE

PRICE
OF
ITEM

NUMBER
OF ITEMS
ORDERED

PURCHASE
ORDER
NO.

RESTOCK
RECORD

ITEM
NO.
CODE

UN IT
CODE

COST
PER
UNIT

MINIMUM
STOCK
QUANTITY

PRESENT
AMT.
OF STOCK

REORDER
NUMBER
11-1520

5-4

Improving SORT Efficiency

Output Data Flies

The output file contains all sales records in the order shown in Table 5-1.

Table 5-1: Sorted Output File
Minor Key: Quantity

Major Key: Item Number

~----i

Lowest quantity

Lowest item no.

Next higher quantity

Highest quantity

Lowest item no.
~-----------Next higher item no.

----------------!
Lowest quantity
Next higher quantity

Next higher item no.

Highest item no.

______________

.,. Highest quantity

-------------~

Lowest quantity
Next higher quantity

Highest item no.

Highest quantity

5.1.1.3 Address Sort - Address sort produces address files, which consist of
record's file addresses (RFAs), beginning at 1, and written in binary words.
These files can be used as a special index file to access randomly the data in
the original file. It is possible to maintain only one data file, but several
different index files as needed. Like tag sort, this process uses the minimum
amount of data necessary in the sorting process. Once the input phase is
completed, the input file is not read again. This means that address sort is the
fastest sorting method of the four SORT types.

NOTE:
Do not transfer an address index file to a device that cannot
handle binary data, such as a printer or terminal.

Improving SORT Efficiency

5-5

The address sort produces an output file consisting of record indices. Each
record index occupies one 6-byte record in the output file. Assume that you
are sorting a file consisting of. six records using the address sort process. If the
sequence of record indices corresponding to the sorted records is 5,1, 6,3,4,2
then the output file can be represented as shown in Figure 5-3.

Figure 5-3: Sample Address Sort Output File
•..

Record's Fiie Address (RFA)

·-

Block Number

RECORD
NUMBER

LOW
(2 bytes
16 bits)

HIGH
(2 bytes
16 bits)

BYTE-IN-BLOCK
(2 bytes
16 bits)

000001

000000

000162

000001

000000

000001

000000

000236

000001

000000

000042

000001

000000

000132

000001

000000

Number of Blocks per File

.__.....

000026
·-"""""

Number of Contiguous
Bytes per Block

Note:
Byte and Block numbers are shown here in hexadecimal, they are written to the actual output
file in binary.

5.1.1.4 Index Sort - Index sort produces an address file consisting of records
file addresses (RF As) in binary, and key fields in original form. This makes it
slightly slower than address sort. During processing this sort handles only the
RF As and two forms of the key fields. One form is used for sorting and the
other is left as it was in the original data.

Index sort p:coduces an output file consisting of record indices plus keys in
original form. Each record in the output file consists of a 6-byte record index
plus the key field.
NOTE:
Do not transfer an index sort output file to a device that cannot
handle binary data, such as a printer or terminal.

5-6

Improving SORT Efficiency

Assume that you are sorting a file consisting of six records using Index sort
process, and you are using a key size of four characters (bytes). The sequence
of record indices corresponding to the sorted record, is 5,1,6,3,4,2 as shown in
Figure 5-4.

Figure 5-4: Sainple Index Sort Output File
Record's Fiie Address 1RFA1
Block Number
RECORD
NUMBER

LOW
(2 bytes
18 bits)

HIGH
(2 bytes
16 bits)

BYTE-IN-BLOCK
(2 bytes
16 bits)

000001

000000

000162

000001

000000

000001

000000

000236

000001

000000

000042

000001

000000

000132

000001

000000

000026

KEY IN ORIGINAL FORM

Notes:
1. ABCD represents the sorting keys In original format.
2. Byte and Block numbers are shown here In hexadecimal, they are written to the actual
output file In binary.

5.1.2 Internal Organization
SORT operates in eight phases (phase 0 through 7). Figure 5-5 summarizes
these phases.
Phase

Function

Decode command line and specification file

Initialize SORT

Get records

Sort records

Initialize merge

lVlerge records

Output records

End SORT

Improving SORT Efficiency

5-7

The VAXNMS command interpreter calls the SORT utility at its main entry
point and Phase 0 is initiated. The initial process statistics are acquired from
the system and stored in a table. SORT calls the command interpreter to
parse and validate the command line. Then SORT validates this information
and stores it in various tables and buffers. If a specification file is present it is
opened, the records are read, the information validated and stored in various
tables and buffers, and the file is closed. Any errors up to this point are
reported by signaling the command interpreter. SORT opens the input file,
and creates the output file.
SORT begins Phase .1. The sorting process is initialized by filling in the key
comparison information, allocating the space needed for input and output
buffers and the sort tree, creating the work files and initializing the sort tree.
Phase 2 begins the sort proper. SORT either reads records from the input file,
or receives them from the caller.
At Phase 3, SORT builds the key from the record and inserts each record into
the tree by key. This process repeats until the sort tree is full or there are no
more records. SORT then outputs the records to the work files as a variable
number of strings each of which is a set of sorted records. Each time a record
is output from the tree a new one is input until there are no more records. The
rest of the records in the tree are output and that ends the initial sorting phase
(phases 1 through 3).
Phase 4 starts SORT's internal merging operation. The memory is redivided
at this point for the merge phase into one to ten input buffers and one output
buffer, depending on the number of initial strings. A different string is read
into each input buffer and the records are merged together into one string and
output to a work file. This process is repeated until the total number of strings
is less than ten.
Phase 5 performs the final merge pass and outputs the remaining string of
records, which is the final sorted file, to either the output file or the caller.
Phase 6 closes the input and output files, closes and deletes the work files, and
returns the me·mory.
Phase 7 acquires the final statistics and prints them, then exits SORT back to
the VAXNMS command interpreter.
Notes:
1. Phases 0 and 7 are part of the utility only.
2. The last part of phase 0 (opening the input file and creating the output
file) and phase 6 are used only by the file 1/0 subroutine package and the

utility.
3. Phases 1through5 are used by the utility and both the file 1/0 and record
1/0 subroutine packages.

5-8

Improving SORT Efficiency

4. Errors during phases 1 through 6 are signaled to the VAXNMS command
interpreter if the utility is running, or returned as a status code to the
caller if the subroutine package is running.
5. All signaled errors produce messages at the command interpreter level
(see Chapter 4).

Figure 5-5: VAX-11 SORT Operating Phases
PHASE 0
DCL SORT Command (via VAX/VMS command Interpreter).
Entry point to SORT.
Get Initial process statistics.
Call command line processor to decode command line.
If It was specified, call specification file decoder.
Open the Input file and create the output flle.
PHASE 1
Set up the key comparison buffer and validate key Information.
Get memory for sorting Initial phase, Input and· output buffers, and set up to read Input.
Create work flies and lnltlallze the sort.
PHASE 2
Read or get record from user and build key.
PHASE 3
Insert record by key Into sort tree.
Is sort tree full? (YES, continue /NO, go back to phase 2)
Output records to work flies as a number of strings of sorted records.
Output and Input until end of file and all records are output.
PHASE 4
Read In strings from work file and merge them until there are 10 or fewer strings left In the
work flies.
PHASE 5
Set up to output records to user or output file.
Do final merge pass outputting records to user files, not work files.
PHASE 6
Delete work files, return memory, close output and Input flies.
PHASE 7
Print statistics and exit.

Improving SORT Efficiency

5-9

5.1.3 Buff'er Allocation and Work Areas
The SORT utility and subroutine package are initially linked with a minimum of space allocated. When SORT is initialized, the work area manager
assigns as much virtual memory as the process needs, and adjusts the working
set size to the process maximum. This allows SORT to minimize page faults
during the sorting, and maximize the order of the merge. At the end of the sort
operation the limits are restored to the size they were at the entry to SORT,
returning the additional virtual memory to the system. SORT requires a
minimum of 75 pages of memory for the working set.
The VAXNMS memory management system service, create and map sec_tion,
allows a user to specify that a particular span of virtual addresses in the
program should be read from and written into a particular set of virtual blocks
within a file on disk, when referenced or paged-out by the entry of another
page. The actual 1/0 to and from the disk is all handled by the pager. SORT
maps the individual virtual addresses representing the work area onto specific
blocks within a work file. When a particular buffer is then referenced within
the work area, the pager automatically brings the correct blocks from the work
file into real memory, and writes the existing blocks back to the correct place
in the file.

5.1.4 Dynamic Memory Usage
Figure 5-6 shows the total address space used by SORT during each of its
eight phases of operation.

Figure 5-6: SORT Dynamic Memory Usage
A= User Limit.
B= User Quota that· must
be greater than or
equal to 75 pag1~s of
real memory.
TAS= Total Address Space.
VM= Virtual Memory.
WS= Working Set.

'r-------,
Unused
Portion of
Code and
Data.

TAS=WS+(=B) "~
VM(=3•B=3•WS)

Code and
Data (fixed)

Phases_..

I
I

I
:

I
I

I
I
r

VM=3•WS(=B)

lvM=
1Buffers(=3•WS=B)I
I
J.
I

Tree=1/2WS
Input Buffer=
1/4WS

Input Buffer
#1

~Same as

Phases 1,2,3

Portion of Code
and Data needed
for Phases 1-3
WS=B

Code and
Data

1,2,3

4,5

__ _B~f~r___ I
1
I Input Buffer I
I
#10
I
,-- - - r----1
~

Output
Buffer

WS=A
TAS=A

r--------1
Output
1

WS=B

Code and
Data (fixed)

WS=A
TAS=A

6,7

._1

1--------~-----~~~~~---SORTRUN-------------------------~..

H-MK-00025-00

5-10

Improving SORT Efficiency

5.1.5 1/0 Considerations
The input and output files 1/0 and the specification file 1/0 are all performed
under the control of VAX-11 RMS record 1/0 facilities. Multi-buffering is
used together with read ahead on unit record devices to optimize the 1/0
operations. The work files are processed by the VAXNMS memory management system service, create and map section.
Various devices can be used for input and output files. Figure 2-1 shows which
devices are allowed for each of the four sorting techniques. Use Figure 2-1 to
match the sorting process with the devices that best suit your processing
environment. Data may be stored in binary, ASCII, decimal, packed or zoned.

5.2 Tuning Procedure
All generalized sorts consider several factors such as: the memory environment (large, small, virtual memory capability or not); the 1/0 devices to be
used for work files and their characteristics (speed, arm movement, seek time,
public or private units); type of files and data most likely to be sorted (large or
small files, large or small records and keys, random or ordered partially,
characters or numbers).
The algorithm must be very good for the cases occurring most often, and
reasonable on all other cases. VAX-11 SORT is designed for an environment
of: fairly large files 1 virtual memory capability, random access disk devices,
public and private, larger random character data files, medium size records
and keys.
There are three components of a sort that account for the majority of the
processing time:
• The number of key comparisons per record per sort.
• The number of merge passes needed to complete the sort.
• The amount of time spent waiting for and/or doing 1/0 to work files.

5.2.1 User Performance Considerations
This section discusses how you can determine the most efficient values for the
following SORT performance parameters.
• Working set quota
• Work file devices
• Number of work files
• Type of sort (process)
• File size

Improving SORT Efficiency

5-11

• Record size
• Key size
• System load
• System process parameters

5.2.1.1 Working Set Quota - For SORT to work efficiently the most important parameter is the working set quota (or size) the user decides to choose.
The optimum working set quota is the smallest one for which the data can be
sorted in memory, that is with 0 merge passes.

To compute the appropriate working set quota size, perform the following
procedure:
Step 1.

For any sort, take the size of the key fields added together in bytes.

Step 2.

Then add 20.

Step 3.

If the sorting process is record sort, add the number of bytes in the

longest record; otherwise add 6. Then multiply by the number of
records in the file. This is the total amount of data you have in
bytes.
Step 4.

Divide that number by 512 to get the amount in blocks.

Step 5.

Multiply by 2 to get the size of the working set quota you should
start with.

For most larger files the number computed will.be much to large to actually
use as a quota. In such cases, the largest reasonable size based on the system
load and scheduling considerations is the correct size to use. An individual
user's authorized quota is generally the largest reasonable size for the
particular system.
For example:
To sort a 1000 record file with 80-byte records and a total key field size of
80 bytes using the record process, compute the following:
1000 X (80+20+80) = 180,000 bytes of data
180,000/512 = 352 blocks of data
352 X 2 = 704 block working set quota

Answer: start with a working set quota of 700.
However, if the same type of a file contained 40,000 records, the total amount
of data would be 14,063 blocks. For most systems a quota of 28,000 blocks
(pages), or even 14,063 is unacceptable. Here the largest reasonable quota
should be used; for example 1024 pages.

5-12

Improving SORT Efficiency

5.2.1.2 Work Fiie Devices ~ Another important parameter is where the work
files are placed. The fuller the disk and the more activity on the disk containing the work files, the less efficient SORT will be. The optimum configuration
would be to have each work file, and the input and output file all on separate
empty disks which are only being used by SORT during the sorting process.
However, this is seldom possible, so the next best configuration is to place
work files on available disks having the lowest activity. See Section 2.7, Setting up the Work Files.

5.2.1.3 Number of Work Flies - Because SORT does not depend on the number of work files used to determine the order of the merge like SORT-11, the
advantage of using more than the default number of work files is limited.
There are two reasons for using more than the default of two work files: 1) to
spread the work files between more than two disks and/or 2) to have each
individual file be a smaller size in order to fit onto a smaller or fuller disk.

If you are using three or four disks, it will help the sort performance to use
three or four work files, one on each disk as discussed above. For example, if
you have a 100,000 block file to sort, using two work files would create two
150,000 block files. But, using four work files would create four 75,000 block
files that could be placed on disks with less free space.

5.2.1.4 Type of Sort - Although the type of sort used is often dictated more
by functionality. required than performance there are significant differences
between the sorts.

Address sort is the fastest and uses the least temporary disk space.
Index sort is only slightly slower than address sort but uses more tempory
storage.
Tag sort uses the same temporary storage as address sort, but is significantly
slower. For large records with small keys it is faster than record sort in smaller
memory sizes if the file is not large.
Record sort uses a larger amount of temporary storage and is the slowest.

5.2.1.5 Using SORT's Statistics -Analyze the sort statistics (Section 2.2.3) to
determine how to improve the sort's performance. The number of records in,
out, sorted if not all equal indicates that there were input or output errors, or
that there are null records in the file (that is, the number of records read was
greater than the number sorted or the number output). This condition can
also be caused by some records containing invalid data in the key fields (if less
than ten records are in error SORT will continue, otherwise SORT will stop
executing).
Longest record length value is obtained from either RMS or the user and can
be used to make sure the RMS value is correct.

Improving SORT Efficiency

5-13

The multl block and buffer counts indicate the amount of 1/0 optimization on

the input and output file. The larger the working set quota the more optimization possible. No optimization would show all these counts as 1. This should
not occur unless the file is huge compared to the working set quota. If it does,
raise the quota if possible.
The order of the merge is the number, less one, of merge buffers that the
working set is divided into for the merge phase.
Number of merge passes and the number of lnltlal runs shows you how close
the data is to fitting in memory. The higher these numbers are, particularly
the number of passes, the longer SORT takes and the further away the working set size is from containing the data.
Virtual memory added is the amount of virtual memory SORT used for the

data.
Elapsed time is the total wall clock time in hours, minutes, seconds, and 1/100
seconds from start to end for the sort run.
The total of the two 1/0 counts are the number of disk hits to get and write data
and these will be higher if the multi block and buffer counts are lower. The
lower the better.
CPU time is the time spent actually processing data minus all 1/0 time. The
closer to the elapsed time the better optimization you are seeing in 1/0.

Page faults are also a good indication of how well the data did or did not fit
into memory. The higher the number of page faults, the less efficient the sort
is.

5.2.2 System Manager Performance Considerations
The system manager can determine the following SORT performance parameter values based on the overall system usage: number of users, types of process
most commonly run, and the amount of real memory available.
• System per process working set quota {WSMAX)
• System per process virtual page count (VIRTUALPAGECNT)
• System per process section count (PROCSECTCNT)
• System modified page writer cluster factor (MPW_ WRTCLUSTER)
The values recommended are based solely on sort considerations; it is up to
the system manager to integrate other system considerations with these in
determining the appropriate final values.

5-14

Improving SORT Efficiency

5.2.2.1 Working Set Quota -The maximum for this value should be set to the
largest size any sort job would ever require. For very large files, working sets of
500 to 1000 pages are not at all unreasonable, provided the system has enough
physical memory to accommodate them. Individual maximums, to prevent
users from monopolizing real memory, can be set on a per user basis by using
the authorization file. For information on how to determine an appropriate
working set for a particular sort job see Section 5.2.1.1. The general rule is, the
smaller the working set, relative to the files to be sorted, the slower the sort.
5.2.2.2 · Vlrtual Page Count - For this parameter the current value as well as
the maximum value should be set to a minimum of 3 to 4 times the value of
the working set quota maximum. When SORT initially starts executing it will
request 2 and 1/2 to 3 times the working set quota of virtual memory from the
system. If this value is too low SORT will be unable to run in certain cases.
5.2.2.3 Process Section Count - For working set quota maximums of 500 or
less this parameter may stay at a minimum level. However, for working set
quotas greater than 500 to 1000 a current value of 10 or greater is necessary. If
this parameter is set too low, SORT will be unable to run in larger working
sets due to internal mapping failures. The value should be increased as the
working set quota maximum increases.
5.2.2.4 Modified Page Writer Cluster Factor -The value of this parameter will

never cause SORT to fail, however it can cause a large difference in performance. For any larger sorts (that is, using working sets of 250 pages or greater)
the larger this parameter, the better. Values of 64 and up are not too large. Be
sure to adjust MPW_HILIM and MPW_LOLIMIT·accordingly. For more
information refer to the SYSGEN procedures in the VAX-11 Software Instal-

lation Guide.

Improving SORT Efficiency

5-15

Glossary

Alphanumeric Characters

The entire set of 128 ASCII characters (see Appendix B).
ASCII Character Set

The set of 128 eight-bit American Standard Code for Information Interchange characters (see Appendix B).
Batch

A mode of processing in which all commands to be executed by the operating system
and, optionally, data to be used as input to the commands are placed in a file or
punched onto cards and submitted to the system for execution.
BLISS

A high-level system implementation programming language. VAX-11 SORT is written in BLISS.
Block

The smallest addressable unit of data that the specified device can transfer in an 1/0
operation (512 contiguous bytes for most disk devices).
Bucket

See File Bucket.
Buffer

A temporary data storage area in a process address space used when performing input
or output operations.
Byte

The smallest addressable unit of information; eight bits. For example, an ASCII
character requires a single byte (see Appendix C for further definitions).
Call

The operation of invoking a procedure.
Caller

The procedure that invoked this procedure by a Call. At the time of procedure
invocation, the invoking procedure is said to be the caller, and the invoked procedure
is the callee. Contrast with User.

Glossary-1

Character

The smallest addressable unit of usable data (byte). It is also a single letter, numeral,
punctuation mark, or other symbol (such as $ or %) , and is represented within the
computer as a unique combination of bits. Typically, a character code consists of
eight bits.

Character String Descriptor

A quadword data structure used for describing character data (strings). The first
word of the quadword contains the length of the character string. The second word
can contain type information. The remaining· longword contains the address of the
string.

CPU

The Central Processor Unit portion of a computer system.
Collating Sequence

The order into which characters are sorted based upon numeric values assigned to
each.
Command

An instruction, generally an English word, typed by the user at a terminal or included in a command. file, which requests the software monitoring a terminal or
reading a command file to perform some well-defined activity. For example, typing
the SORT command request the system to invoke the SORT utility.
Command Fiie

A file containing command strings. See also Command Procedure.
Command Interpreter

Procedure-based system code that executes in supervisor mode in the context of a
process to receive, syntax check, and parse commands typed by the user at a terminal
or submitted in a command file.
Command Parameter

The positional operand of a command delimited by spaces, such as a file specification, option, or constant.
Command Procedure

A file containing commands and data that the command interpreter can accept in
lieu of the user typing the commands individually on a terminal.

Glossary-2

Command String

A line (or set of continued lines), normally terminated by typing the carriage return
key, containing a command and, optionally, information modifying the command. A
complete command string consists of a command, its qualifiers, if any, and its
parameters (file specifications, for example), if any, and their qualifiers, if any.
Compatlblllty Mode

A mode of execution that enables the central processor to execute non-privileged
PDP-U instructions. The operating system supports compatibility mode execution
by providing an RSX-UM programming environment for an RSX-UM task image.
The operating system compatibility mode procedures reside in control region of the
process executing a compatibility mode image. The procedures intercept calls to the
RSX-UM executive and convert them to the appropriate operating system functions.
Contiguous Blocks

Physically adjacent and/or consecutively numbered blocks of data.
Data Fiie Record

A record containing user data.
Data Structure

Any table, list, array, queue, or tree whose format and access conventions are well
defined for reference by one or more images.
Data Type

In general, the way in which bits are grouped and interpreted. In reference to the
processor instructions, the data type of an operand identifies the size of the operand
and the significance of the bits in the operand. Operand data types include: byte,
word, longword, and quad-word integer, floating and double floating, character
string, packed decimal string, and variable-length bit field (see Appendix C).
DCL

Digital Command Language (DCL) is a set of English- like statements that a user
types to initiate and control system operations.
Default

An assumed value supplied to the system when a command qualifier does not specifically override the normal command function; fields in a file specification that the
system fills in when the specification is not complete.
Descriptor

See Character String Descriptor.

Glossary-3

Device

The general name :for any physical terminus or link connected to the processor that is
capable of receiving, storing, or transmitting data. Card readers, line printers, and
terminals are examples of record-oriented devices. Magnetic tape devices and disk
devices are examples of mass storage devices. Terminal line interfaces and interprocessor links are examples .of communications devices.
Directory

A file used to locate files on a volume that contains a list of file names (including
extension and version number) and their unique internal identifications.
Directory Name

The field in a file specification that identifies the directory file in which a file is
listed. The directory name is enclosed in brackets ( [] or <>).
Fie Id

A logically distinguishable area within a record. Usually a logical unit of data.
Fiie

A logically related collection of data on a volume such as disk or magnetic tape. A file
can be referenced by a name assigned by the user. A file normally consists of one or
more logical records.
Fiie Bucket

Within the RMS Relative File organization, a bucket is a storage structure of one to
32 blocks of data.
Fiie Header

A block in the index file describing a file on a FILES-11 disk structure. The file
header identifies the locations of the file's extents. There is a file header for every file
on the disk.
Fiie Organization

The particular file structure used to record the data constituting a file on a mass
storage medium. RMS file organizations are: Sequential, Relative, and Indexed.
Fiie Prologue

The first block in a relative or indexed file which contains header information for the
file.
File Specification

A unique name for a file on a mass storage medium. It identifies the node, the device,
the directory name, the file name, and the version number under which a file is stored
(see Appendix D for additional information).

Glossary-4

Fiie Structure

The way in which the blocks forming a file are distributed on a disk or magnetic tape
to provide a physical accessing technique suitable for the way the data in the file is
processed.
Fiie System

A method of recording, cataloging, and accessing files on a volume.
Fiie Type

The field in a file specification that is preceded by a period or dot(.) and consists of a
zero-to three-character type identification. By convention, the type identifies a generic class of files that have the same use or characteristics, such as ASCII text files,
binary object files, etc.
Flles-11

The standard physical disk structure used by VAX-11 RMS.
Flies pee

Fi.le Specification that uniquely identifies a file by physical location (see Appendix
D).

Fiie, Input

See Input File.
Fiie, Output

See Output File.
Fiie, Work

See Work File.
Fixed Control Area

An area associated with a variable length record available for controlling or assisting
record access operations. Typical uses include line numbers and printer format control information.
Fixed Position Fleld

An area associated with character position (or column numbers). Used in SORT-11
Specification Files.
Fixed Length Record Format

A file format in which all records have the same length.

Glossary-5

Format

The arrangement of any record or file; the order in which fields reside in a record.
Free Fields

Logically positioned fields separated by commas. Contrast with fixed position fields.
Home Block

A block in the index file that contains the volume identification, such as volume label
and protection.
Image

A file consisting of procedures and data that have been bound together by the linker.
There are three types of images: Executable, Shareable, and System.

Indexed Fiie Organization

A file organization in which a file contains records and a primary key index (and
optionally one or more alternate key indices) used to process the records sequentially
by index or randomly by index.
Index Fiie

The file on a Files-11 volume that contains the access information for all files on the
volume and enables the operating system to identify and access the volume.
Index Fiie Bit Map

A table in the index file of a Files-11 volume that indicates which file headers are in
use.
Index Fiie Record

A record of file system data that is invisible to the user.
Input File

The file containing the records you wish to sort.
Key, Key Fie Id

The data field containing the values chosen from a record to control the sort (see
section 2.6).
Key, Major

The most important field in the total key. If you were sorting a list by department,
salary and name, department would be the major key.

Glossary-6

Key, Minor

The least significant field in the total key. In the preceding example name is the
minor key.
Library

A collection of commonly used files.
Line

In this document, a line generally refers to a line in the SORT specifications form or a
record in the specification file.
Logical Name

A user-specified name for any portion or all of a file specification. For example, the
logical name INPUT can be assigned to a terminal devic_e from which a program
reads data entered by a user. Logical name assignments are maintained in logical
name tables for each process, each group, and the system. A logical name can be
created and assigned a value permanently or dynamically.
Longword

Four contiguous bytes (32 bits) starting on an addressable byte boundary (see Appendix C).
LRL

Longest Record Length (LRL) specified in bytes.
Merge

A process by which two or more ordered groups of records are put together record-byrecord into a single identically ordered group.
Native Mode

The processor's execution mode, in which the programmed instructions are interpreted as byte-aligned, variable length instructions that operate on byte, word, longword, quadword integ·er, floating and double floating, character string, packed decimal, and variable length bit field data. The instruction execution mode other than
compatibility mode.
Node

An individual computer system in a network.
Output Fiie

The file created by running SORT. The output file may be either a data file or an
address file.

Glossary-7

Page

1). A set of 512 contiguous byte locations used as the unit of memory mapping and
protection. 2). The data between the beginning of file and a page mark~r, between
two markers, or between a marker and the end of a file.
Page Fault

An exception generated by a reference to a page which is not mapped into a working
set.
Pager

A set of kernel mode procedures that executes as the result of a page fault. The pager
makes the page for which the fault occurred available in physical memory so that the
image can continue execution. The pager and the image activator provide part of the
operating system's memory management functions.
Paging

The action of bringing pages of an executing process into physical memory when
referenced. When a process executes, all of its pages are said to reside in virtual
memory. Only the actively used pages, however, need to reside in physical memory.
In this system, a process is paged only when it references more pages than it is
allowed to have in its working set. When a process refers to a page not in its working
set, a page fault occurs. This causes the operating system's pager to read in the
referenced page fault if it is on disk (and optionally, other related pages depending on
a cluster factor), replacing the least recently faulted pages as needed. This system
only pages a process against itself.
Packed Decimal

A method for compact storage of numeric data; two digits are stored in each 8-bit
byte and the sign resides in the last byte of the low-order digit.
Parse

To break down into individual parts from a whole.
Physical Memory

The memory modules contained within the CPU. Also called main memory.
Procedure

A routine that follows the VAX-11 calling sequence standard. A procedure may
return values via the argument list and/or the standard value return registers. Contrast with routine.
Process

The basic entity scheduled by the system software that provides the context in which
an image executes. A process consists of an address space and both hardware and
software context.

Glossary-8

Program

A program is the basic entity that is executed by the processor. Each program consists of a set of procedures and its execution represents a distinct activity that is
potentially concurrent with others in the system.
Prologue

See File Prologue.
Quadword

Eight contiguous bytes (64 bits) starting on an addressable byte boundary. See Appendix C.
Quallfler

A portion of a command string that modifies a command verb or command parameter by selecting one of several options. A qualifier, if present, follows the command
verb or parameter to which it applies and is in the format: "/qualifier=option". For
example, in the command string "PRINT filename/ COPIES=3", the COPIES qualifier indicates that the user wants three copies of a given file printed.
Random acces~ by record's flle address

The retrieval of a record by its unique address, which is provided to the program by
RMS. The· method of access can be used to randomly access a sequentially organized
file containing variable length records.
Random access by relative record number

The retrieval or storagE~ of a record by specifying its position relative to the beginning
of the file.
Real Memory

See Physical Memory.
Record

The unit of information in a file; a group of related fields treated as a logical unit.
Record Cell

A fixed length area in a relatively organized file that is used to contain one record.
Record Management Services (RMS)

A set of system procedures in the operating system that are called by programs to
p'rocess files and records within files. RMS allows programs to issue GET and PUT
requests at the record level (record 1/0) as well as read and write blocks (block 1/0).
RMS is an integral part of the system software. RMS procedures run in executive
mode.

Glossary-9

Record-oriented Device

A device such as a terminal, line printer, or card reader, on which the largest unit of
data that a program can access is the device's physical record.
Record's Fiie Address (RFA)

The unique address of a record in a file that allows records to be accessed randomly
regardless of file organization.
Record, Data Fiie

See Data File Record.
Record, Field Specification

See Section 2.5.2
Record, Header

See Section 2 .5 .2
Record, Index Fiie

See Index File Record.
Relative Fiie Organization

A file organization in which the file contains fixed length record cells. Each cell is
assigned a consecutive number that represents its position relative to the beginning of
the file. Records within each cell can be the same length or smaller than the cell.
Relative file organization permits sequential record access, random record access by
record number, and random record access by record's file address.
RMS

See Record Management Services (RMS).
Routine

A sequence of instructions that performs a well defined action. It may have multiple
entry points. For example, the SIN routine has SIN and COS entry points. A routine
that follows the VAX-11 calling sequence standard is termed a procedure.
Sequential Fiie Organization

A file organization in which records appear in the order in which they were originally
written. The records can be fixed length or variable length. Although one does not
speak of record cells with sequentially organized files, for purposes of comparison
with relatively organized files one can say that the record itself is the same as its
record cell, and its record number is the same as its relative cell number. Sequential
file organization permits sequential record access and random record access by
record's file address. Sequential file organization with fixed length records also permits random access by relative record number.

Glossary-IO

SORT Utlllty

A processing program that can be used to sort records by keys into a prescribed
sequence. To segregate items into groups according to some definite rules.
Sort Tree

A data structure used to keep order of records by sort.
Subroutine

A procedure that does not return a known value in the value registers. If values are
returned, they are returned via the argument list. By convention, the function "value"
is unpredictable.
System Device

The device on which the Executive resides.
Terminal

The general name for those peripheral devices that have keyboards and video screen
or printers. Under program control, a terminal enables people to type commands and
data on a keyboard and receive messages on a video screen or printer. Examples of
terminals are the LA3H DECwriter (hard-copy terminal) and the VT52 video display
terminal (soft-copy terminal).
Unit Record Device

See Record-oriented Device.
User

The person who is directly using the computer, either via terminal or batch input.
Contrast with Caller.
Variable-length Record

A record format in which records need not be the same length.
Variable with fixed-length control (VFC) record

A record format in which records of variable length contain an additional fixed-length
control area. The control area may be used to contain file line numbers and/or print
format control characters.
Virtual Address

A 32-bit integer identifying a byte "location" in virtual address space. The memory
management hardware translates a virtual address to a physical address. The te'rm
virtual address may allso refer to the address used to identify a virtual block on a mass
storage device.

Glossary-11

Virtual Memory

The set of storage locations in physical memory and on disk that are referred to by
virtual addresses. From the programmer's viewpoint, the secondary storage locations
appear to be locations in physical memory. The size of virtual memory in any system
depends on the amount of physical memory available and the amount of disk storage
used for non-resident virtual memory.
Volume

A mass storage medium such as a disk pack or reel of magnetic tape.
Wiid card

The asterisk character when used as a substitute parameter in file specification
indicates "all" for a given field.
Word

Two contiguous bytes (16 bits) starting on an addressable byte boundary (see Appendix C).
Work Fiie

A collection of sortE~d records created during the processing cycle and released after
the sort is finished. (Sometimes called Scratch Files.)
Working Set

The set of pages in process address space to which an executing process can refer
without incurring a page fault. The working set must be resident in memory for the
process to execute. The remaining pages of that process, if any, are either in memory
and not in the process working set or they are on secondary storage.
Zoned Numeric Format

A specific ASCII coded decimal data type where the number sign and the least
significant digit are combined into a single hexadecimal code (see Appendix C).

Glossary-12

Appendix A
Octal/Hexadecimal/Decimal Conversion

A.1 Octal/Declmal Conversion
To convert a number from octal to decimal, locate in each column of the table
the decimal equivalent for the octal digit in that position. Add the decimal
equivalents to obtain the decimal number.
To convert a decimal number to octal:
1. locate the largest decimal value in the table that will fit into the decimal

number to be converted,
2. note its octal equivalent and column position,
3. find the decimal remainder.
Repeat the process on each remainder. When the remainder is 0, all digits will
have been generated.

32,768

4,096

512

65,536

8,192

1,024

128

98,304

12,228

1,536

192

31,072

16,384

2,048

256

163,840

20,480

2,560

320

16£1,608

24,576

3,072

384

22£1,376

28,672

3,584

448

A-1

A.2 Powers of 2 and 16
Powers of 16

Powers of 2

2**n

16**n

256
512
1024
2048
4096
8192
16384
32768
65536
131072
262144
524288
1048576
2094304
4194304
8388608
16777216

8
9
10

1
16
256
4096
65536
1048576
16777216
268435456
4294967296
68719476736
1099511627776
17592186044416
281474976710656
4503599627370496
72057594037927936
1152921504606846976

0
1
2
3
4

12
13
14
15
16
17
18
19
20
21
22
23
24

6
7
8
9
10
11

12
13
14
15

A.3 Hexadecimal to Decimal Conversion
For each integer position of the hexadecimal value, locate the corresponding
column integer and record its decimal equivalent in the conversion table A.5.
Add the decimal equivalent to obtain the decimal value.
Example:
D0500ADO Cl G >

?< 10)

00000000
500000
AOO
DO

=
=
=
=

3,499,sso,920
5,zaz,sso
2 , 580
208

D0500ADO

3,a9a,9oa,s7s

A.4 Decimal to Hexadecimal Conversion
1. Locate in the conversion table A.5 the largest decimal value that does not

exceed the decimal number to be converted.
2. Record the hexadecimal equivalent followed by the number of zeros (0)
that corresponds to the integer column minus one.
"

3. Subtract the table decimal value from the decimal number to be
converted.
4. Repeat steps 1-3 until the subtraction balance equals zero (0). Add the
hexadecimal equivalents to obtain the hexadecimal value.

A-2

Octal/Hexadecimal/Decimal Conversion

Example:

=
=
=
=
=
=
=

22t4GG ( 1 0 )
20t480
1 t792
192
2

------

22t4GG

? ( 1 G>

5000
700

22t4GG
-20t480

1 t98G
- 1 t 792

57C2
19£1
192
2
2
0

A.5 Hexadecimal Integer Columns

8
DEC

HEX

0
1
2
3
4
5
6
7
8
9

0
268,435,456
536,870,912
805,306,368
1,073,741,824
1,342, 177 ,280
1,610,612, 736
1,879,048,192
2,147,483,643
2,415,919, 104
2,684,354,560
2,952, 790,016
3,221,225,472
3,489,660,928
3, 758,096,384
4,026,531,840

0
1
2
3
4
5
6
7
8
9

A
B

c
D

E
F

HEX

A
B

c
D

E
F

DEC

0
1,048,576
2,097,152
3,145,728
4,194,304
5,242,880
6,291,456
7,340,032
8,338,608
9,437,184
10,485,760
11,534,336
12,582,912
13,631,488
14,680,064
15,728,640

0
16,777,216
33,554,432
50,331,648
67,108,864
83,886,080
100,663,296
117 ,440,512
134, 217' 728
150,994,944
167,772,160
184,549,376
201,326,592
218, 103,808
234,881,024
251,658,240

2
3
4
5
6
7
8
9

A
B

c
D

E
F

0
2
3
4
5
6
7
8
9

E
F

HEX

0
65,536
lill,072
196,608
262, 144
:m,680
::193,216
458,752
fi24,288
fi89,824
. 6fifi,360
720,896
786,432
851,968
917,504
983,040

0
1
2
3
4
5
6
7
8
9

c
D

E
F

BYTE

BYTE
\

WORD

HEX 1 DEC

HEX

DEC

HEX

DEC

HEX

0
1
2
3
4
5
6
7
8
9

0
256
512
768
1,024
1,280
1,536
1,792
2,048
2,304
2,560
2,916
3,072
3,328
3,584
3,840

0
1
2
3
4
5
6
7
8
9

0
4,096
8,192
12,288
16,384
20,480
24,576
28,672
32,768
36,864
40,960
45,056
49,152
53,248
57,344
61,440

A
B

c
D

E
F

BYTE

c
D
E
F

0
16
32
48
64
80
96
112
128
144
160
176
192
208
224
240

0
2
3
4
5
6
7
8
9

E
F

0
1
2
3
4
5
6
7
8
9
10
11

12
13
14
15

BYTE

v
WORD

1
DEC HEX DEC

LONGWORD

Octal/Hexadecimal/Decimal Conversion A-3

Appendix B
The ASCII Chara.cter Set Collating Sequence

ASCII
Character

Hexadecimal
Number

ASCII
8--Bit
Octal

Decimal

ASCII
Character

Hexadecimal
Number

ASCII
8-Bit
Octal

Decimal

NUL

000

034

SOH

001

035

STX

002

036

ETX

003

037

EOT

004

040

ENQ

005

041

ACK

006

042

BEL

007

043

010

044

011

045

012

046

013

047

014

(

050

015

)

051

016

052

017

053

DLE

020

054

DCl

021

055

DC2

022

056

DC3

023

057

DC4

024

060

NAK

025

061

SYN

026

062

ETB

027

063

CAN

030

064

031

065

SUB

032

066

ESC

033

067

(continued on next page)

B-1

·~-..tt.<'<·,jl;;;;o',,

,,~'

Hexadecimal
Number

ASCII
8-Bit
Octal

Decimal

134

135

072

136

073

137

074

140

075

141

076

142

077

143

100

144

100

101

145

101

102

146

102

103

147

103

104

150

104

105

151

105

106

152

106

107

153

107

110

154

108

111

155

109

112

156

110

113

157

111

114

160

112

115

161

113

116

162

114

117

163

115

120

164

116

121

165

117

122

166

118

123

167

119

124

170

120

125

171

121

126

172

122

127

173

123

130

174

124

131

175

125

132

176

126

[

-· 133

DEL

177

127

Hexadecimal
Number

ASCII
8-Bit
Octal

Decimal'

070

071

ASCII
Character

B-2

ASCII
Character

Characters Set ASCII Collating Sequence

Appendix C
Data Types

The data type refers to the way in which bits are grouped and interpreted. In
reference to the processor instructions, the data type of an operand identifies
the size of the eiperand and the significance of the bits in the operand. Data
types applicable to SORT and its associated VAXNMS programs are: separated into threE! classes; character, binary, and decimal. These classes can be
subdivided into data types of different sizes and formats such as; byte, word,
longword, quad.word, floating, double floating, character string, packed decimal string, and variable-length bit field.

C.1 Byte
A byte is 8 contiguous bits starting on an addressable byte boundary. The bits
are numbered from the right 0 through 7:
7

c__]

A byte is specified by its address A. When interpreted arithmetically, a byte is
a twos complement integer with bits of increasing significance going 0 through
6 and bit 7 the sign bit. The value of the integer is in the range -128 through
127. For the purposes of addition, subtraction, and comparison, VAX-11 instructions also provide direct support for the interpretation of a byte as an
unsigned integer with bits of increasing significance going 0 through 7. The
value of the unsigned integer is in the range 0 through 255.

C.2 Word
A word is 2 contiguous bytes starting on an arbitrary byte boundary. The bits
are numbered from the right 0 through 15:
1
5

_ __

A word is specified by its address A, the address of the byte containing bit 0.
When interpreted arithmetically, a word is a twos complement integer with
bits of increasing significance going 0 through 14 and bit 15 the sign bit. The
value of the integer is in the range -32, 768 through 32, 767. For the purposes of

C-1

addition, subtraction and comparison, VAX-11 instructions also provide direct support for the interpretation of a word as an unsigned integer with bits of
increasing significance going 0 through 15. The value of the unsigned integer is
in the range 0 through 65,535.

C.3 Longword
A longword is 4 contiguous bytes starting on an arbitrary byte boundary. The
bits are numbered from the right 0 through 31:
3
0

'-----------·--~

A longword is specified by its address A, the address of the type containing bit
0. When interpreted arithmetically, a longword is a twos complement integer
with bits of increasing significance going 0 through 30 and bit 31 the sign bit.
The value of the integer is in the range -2,147,483,648 through 2,147,483,647.
For the purposes of addition, subtraction, and comparison, VAX-11 instructions also provide direct support for the interpretation of a longword as an
unsigned integer with bits of increasing significance going 0 through 31. The
value of the unsigned integer is in the range 0 through 4,294,967,295.
Note that the longword format is different from the longword format defined
by the PDP--11 FP-11. In that format, bits of increasing significance go from
16 through 31 and 0 through 14. Bit 15 is the sign bit. Most DIGITAL software
and in particular PDP-11 FORTRAN and COBOL use the VAX-11 longword
format.

C.4 Quadword
A quadword is 8 contiguous bytes starting on an arbitrary byte boundary. The
bits are numbered from the right 0 through 63:
3

E
6
3

:A
:A+4

A quadword is specified by its address A, the address of the byte containing
bit 0. When interpreted arithmetically, a quadword is a twos complement
integer with bits of increasing significance going 0 through 62 and bit 63 the
sign bit. The value of the integer is in the range -2**63 to 2**63--1. The quadword data type is not fully supported by VAX-11 instructions.

C-2

Data Types

C.5 Floating
A floating datum is 4 contiguous bytes starting on an arbitrary byte boundary.
The bits are labelled from the right 0 through 31.
1 1 1
6 5 4

fraction

Isl

7 6
exp

L_ -~raction

]

A floating datum is specified by its address A, the address of the byte containing bit 0. The form of a floating datum is sign magnitude with bit 15 the sign
·bit, bits 14:7 an excess 128 binary exponent, and bits 6:0 and 31:16 a normalized 24--bit fraction with the redundant most significant fraction bit not represented. Within the fraction, bits of increasing significance go from 16 through
31and0 through 6. The 8-bit exponent field encodes the values 0 through 255.
An exponent value of 0 together with a sign bit ofO, is taken to indicate that
the floating datum has a value of 0. Exponent values of 1 through 255 indicate
true binary exponents of -127 through +127. An exponent value of 0, together
with a sign bit of 1, is taken as reserved. Floating point instructions processing
a reserved operand take a reserved operand fault (See Chapter 4 and 6). The
value of a floating datum is in the approximate range .29*10**-38 through
1. 7*10**38. The precision of a floating datum is approximately one part in
2**23, that is, typically 7 decimal digits.

C.6 Double Floating
A double floating datum is 8 contiguous bytes starting on an arbitrary byte
boundary. The bits are labelled from the right 0 through 63:
3

1 1 1
6 5 4

fraction

6
3

exp

7 6

fraction
4 4
8 7

:A
:A+4

3
2

A double floating datum is specified by its address A, the address of the byte
containing bit 0. The form of a double floating datum is identical to a floating
datum except for an additional 32 low significance fraction bits. Within the
fraction, bits of increasing significance go 48 through 63, 32 through 47, 16
through 31, and 0 through 6. The exponent conventions, and approximate
range of values is the same for double floating as floating. The precision of a
double floating datum is approximately one part in 2**55, that is, typically 16
decimal digits.

Data Types

C-3

C. 7 Variable Length Bit Fleld

A variable bit field is 0 to 32 contiguous bits located arbitrarily with respect to
byte boundaries. A variable bit field is specified by 3 attributes: the address A
of a byte, a bit position P which is the starting location of the field with
respect to bit 0 of the byte at A, and a size S of the field. The specification of a
bit field is indicated by the following where the field is the shaded area.

P+S

P+S-1

P-1

_=1.:A
S-1

The position is in the range -2**31 through 2**31-1 and is conveniently
viewed as a signed 29-bit offset and a 3-bit bit-within-byte (BWB) field:

3
3 2

--~~------~~~b-y-te_of-fs_e_t___________,____~____I_. __:~~

The sign extended 29-bit by~e offset is added to the address A and the resulting address specifies the byte in which the field begins. The 3-bit bit-withinbyte field encodes the starting position (0 through 7)· of the field within that
byte. The VAX-11 field instructions provide direct support for the interpretation of a field as a signed or unsigned integer. When interpreted as a signed
integer, it is twos complement with bits of increasing significance going 0
through S-2; bit S-1 is the sign bit. When interpreted as an unsigned integer,
bits of increasing significance go from 0 to S-1. A field of size 0 has a value
identically equal to 0.

A variable bit field may be contained in 1to5 bytes. From a memory management point of view, only the minimum number of bytes necessary to contain
the field is actually referenced.

C-4

Data Types

C.8 Character String
A character string is a contiguous sequence of bytes in memory. A character
string is specified by 2 attributes: the address A of the first byte of the string,
and the length L of the string in bytes. Thus the format of a character string
is:
0

:A+L-1

The address of a string specifies the first character of a string. Thus 11 XYZ 11 is
represented:
"X"

"Y"

:A+1

"Z"

:A+2

The length L of a string is in the range 0 through 65,535.

C.9 Tralllng Numeric String
A trailing numeric string is a contiguous sequence of bytes in memory. The
string is specified by 2 attributes: the address A of the first byte (most significant digit) of the string, and the length L of the string in bytes.
All bytes of a trailing numeric string, except the least significant digit byte,
must contain an ASCII decimal digit character (0-9). The representation for
the high order digits is:
digit

decimal

hex

48
49
50
51
52
53
54
55
56
57

30
31
32
33
34
35
36
37
38
39

2
3
4
5
6
7
8
9

ASCII
character

0
1
2
3
4

5
6
7
8
9

Data Types

C-5

The highest addressed byte of a trailing numeric string represents an encoding
of both the least significant digit and the sign of the numeric string. The VAX
numeric string instructions support any encoding; however, there are 3 preferred encodings used by DIGITAL software. Thes~ are (1) unsigned numeric
in which there is no sign and the least significant digit contains an ASCII
decimal digit character, (2) zoned numeric, and (3) overpunched numeric.
Because the overpunch format has been used by compilers of many manufacturers over many years, and because various card encodings are used, several
variations in overpunch format have evolved. Typically, these alternate forms
are accepted on input. The valid representations of the digit and sign in each
of the later two formats is:
Representation of Least Significant Digit and Sign
Zoned Numeric Format
digit
0
1
2
3

decimal
48
49
50
51
52
53
54
55
56
57
112
113
114
115
116
117
118
119
120
121

5
6
7
8
9
-0
-1
-2
-3
-4
-5
-6
-7
-8
-9

hex
30
31
32
33
34
35
36
37
38
39
70
71
72
73
74
75
76

ASCII
char
0
1
2
3
4
5
6
7
8
9
p

q
r

t
u
v

78
79

x
y

Overpunch Format
decimal
123
65
66
67
68
69
70
71
72
73
125
74
75
76

ASCII char
norm
alt.

hex
7B
41
42
43
44
45
46
47
48
49
7D
4A
4B
4C
4D
4E
4F
50
51
52

78
79
80
81
82

A
B

a
b
c
d
e
f
g
h
i

D
E
F
G

H
I

]!:

J
K
L
M
N
0

k
1
m
n
0

p
q
r

Q
R

The length L of a trailing numeric string must be in the range 0 to 31 (0 to 31
digits). The value of a 0 length string is identically 0. The address A of the
string specifies the byte of the string containing the most significant digit.
Digits of decreasing significance are assigned to increasing addresses. Thus
"123" is represented:
Zoned Format or Unsigned
4 3

C-6.

Data Types

Overpunch Format
4 3

:A+1

:A+2

and "-123" is represented:
Overpunch Format

Zoned Format

:A+1

:A+2

C.10 Leading Separate Numeric String
A leading separate numeric string is a contiguous sequence of bytes in memory. A leading separate numeric string is specified by 2 attributes: the address
A of the first byte (containing the sign character), and a length L, which is the
length of the string in digits and NOT the length of the string in bytes. The
number of bytes in a leading separate numeric string is L+l.
The sign of a separate leading numeric string is stored in a separate byte.
Valid sign bytes are:
Sign

decimal

hex

ASCII character

+
+

43
32
45

2B
20
2D

<blank>

The preferred representation for 11 + 11 is ASCII "+". All subsequent bytes contain an ASCII digit character:
digit

decimal

hex

ASCII character

0
1
2
3

48
49
50
51
52
53
54
55
56
57

30
31
32
33
34
35
36
37
38
39

e:·

{;

1
2
3
4
5
6
7
8
9

The length L of a leading separate numeric string must be in the range 0 to 31
(0 to 31 digits). The value of a 0 length string is identically 0.

Data Types

C-7

The address A of the string specifies the byte of the string containing the sign.
Digits of decreasing significance are assigned to bytes of increasing addresses.
Thus "+123" is: ·
4 3

:A+1

:A+2

:A+3

and "-123" is:
0

7
2

. C.11 Packed Decimal String
A packed decimal string is a contiguous sequence of bytes in memory. A
packed decimal string is specified by 2 attributes: the address A of the first
byte of the string and a length L which is the number of digits in the string
and NOT the length of the string in bytes. The bytes of a packed decimal
string are divided into 2 4-bit fields (nibbles) which must contain decimal
digits except the low nibble (bits 3:0) of the last (highest addressed) byte
which must contain a sign. The representation for the digits and sign is:

C-8

Data Types

digit or sign

decimal

hex

0
1
2
3
4
5
6
7
8

10,12,14, or 15
11 .or 13

A,C,E, or F
B, or D

The preferred sign representation is 12 for 11 + 11 and 13 for 11 - 11 • The length L is
the number of digits in the packed decimal string (not counting the sign) and
must be in th~~ range 0 through 31. When the number of digits is odd, the
digits and' the sign fit in L/2 (integer part only) + 1 bytes. When the number of
digits is even, it is required that an extra 11 0" digit appear in the high nibble
(bits 7:4) of the first byte of the string. Again the length in bytes of the string
is L/2 + 1.
The address A of the string specifies the byte of the string containing the most
significant digit in its high nibble. Digits of decreasing significance are assigned to increasing byte addresses and from high nibble to low nibble within
a byte. Thus 11 +123" has length 3 and is represented:
7

4 3

CGJ:A

~:A+1

and 11 -12 11 has length 2 and is represented:
7

4 3

:A+1

Data Types · C-9

Appendix D
Data Structures and Basic Concepts

This Appendix. provides beginning users with additional information regarding the following topics:
• VAX-11 RMS data files, records, and structures
• File specification parameters
• Programming languages

D.1 Data Flies, Records, and Structures

D.1.1 Data Hierarchy
Four level data hierarchy (character, field, record, file) is shown in Figure D.1:

Figure D.1: Data Hierarchy

m
I

character
I

Adams, Jotii!]J

field

.,,,,..,
,...,

} $175 l
.
.
--- . . ---=

Adams r638-36-100& /

record

-- -

A file is a collection of related information. For example, a file might contain a
company's personnel information (employee names, addresses, and job titles).
Within this file, the information is divided into records. All the information on
a single employee might constitute a single record. Each record in the personnel file would be subdivided into discrete pieces of information known as
fields. By specifying key fields (Keys) in a particular order, you can sort entire
records into any order. Using VAX-11 SORT you can retrieve records in
ascending or descending order by ordered key fields (that is, create sorted
data files); and you can create sorted address files for random record retrieval
by user programs.

D-1

D.1.2 Record Types
VAX-11 SORT processes three diff~rent types of records: Fixed, Variable, and
Variable with fixed-length control (VFC). Figure D.2 summarizes these record
types.

Figure D.2: Record Types (fixed, variable, VFC)
FIXED LENGTH 32-BYTE RECORDS

24
bytes

8
bytes

·------

\..__ _ _ RECORD 1 * - - - - ' \----RECORD 2* _ __./'-.~--RECORD 3* - - *NOTE: VAX-11 RMS considers all 32 bytes to be used, even though they may not contain
useful Information In the eyes of the user.

VARIABLE LENGTH RECORDS ON MAGTAPE
4-Byte Count Fields

I 8
lbytes

16
bytes

2-Byte Count Fields

i:~4
bytes

.........-------~~~--.......

'----./ '----1'RECORD 1

RECORD 2

VARIABLE LENGTH RECORDS ON DIS~·

RECORD 3

....__b_y_te_s__.,....!.

w--___

\..._/\
RECORD 1 RECORD 2

I
RECORD 3

VARIABLE WITH FIXED-LENGTH CONTROL RECORD (VFC)

Count
field

Fixed control area
portion

-===i

'""----RECORD 1 _ _ _ _ _/
F-MK-00026-00

D-2 Data Structures and Basic Concepts

D.1.3 VAX-11 RMS Fiie Organizations

1. Sequential Flies (see Figure D.3) Sequential files may contain the follow-

ing record types:
a. Fixed LEmgth Records
b. Variable Length Records
c. Variable with Fixed-length Control (VFC) Records
The order of the records in a sequential file is determined by the order in
which the records were originally written to the file. The first record in the
file is the first record read; the second record next, and so on.Sequential
files are the only files permitted for magnetic tape and unit record devices.
They are also permitted for disk.
2. Relative Flies - Records may be any type (that is, fixed, variable, or VFC)
as long as the maximum record length is specified. Each record is
numbered 1 ton relative to the beginning of file (as shown in Figure D.4).
A relative file consists of record areas (cells) that are identified by relative
record numbers. The first record area in the file is record number 1, the
second is 2, and so on. Empty or null records are permitted. Relative files
can reside only on disk.
Relative file considerations:
• Most efficient random access in terms of speed and storage space overhead.
• Addresses of records (relative record numbers) must be known to process
file randomly.
• Requires storage space to contain all record positions from record number one to highest record number stored in file.
• Records can span blocks, but cannot span buckets.
• Can be write shared.
3. Indexed FHes - Contain one or more indices, as well as data records.
Records can be of any type (that is, fixed, variable, or VFC) as long as the
maximum :record length is specified. To retrieve information, you ask for
the proper record by primary or alternate key. The system looks up the
key in the appropriate index and retrieves the record using the record
pointer associated with the key. Indexed files can reside only on disk.
The location of records in the indexed file organization is transparent to
the program. RMS completely controls the placement of records in an
indexed fih~. The presence of keys in the records of the file governs this
placement.

Data Structures and Basic Concepts

D-3

A key is a field present in every record of an indexed file. The location and
length of this field are identical in all records. When creating an indexed
file, the user decides which field or fields in the file's records are to be a
key. Selecting such fields indicates to RMS that the contents (that is, key
value) of those- fields in any particular record written to the file can be
used by a program to identify that record for subsequent retrieval.
At least one key must be defined for an indexed file: the primary key.
Optionally, additional keys or alternate keys can be defined. An alternate
key value can also be used as a means of identifying a record for retrieval.
As programs write into an indexed file, RMS builds a tree-structured table
_known as an index. An index consists of a series of entries containing a key
value copied from a record that a program wrote into the file. Stored with
each key value is a pointer to the location in the file of the record from
which the value was copied. RMS builds and maintains a separate index
for each key defined for the file. Each index is stored in the file. Thus,
every indexed file contains at least one index, the primary key index.
Figure D.5 shows an RMS indexed file organization with a primary key.
When alternate keys are defined, RMS builds and stores an additional
index for each alternate key.
Index file considerations:
• Multi-key indexed sequential capability.
• Most flexible in terms of how a record is accessed.
• A record is addressed by the contents of a field in the record (the key
field).
• Records can be retrieved sequentially in a collated order by key field.
• Requires the most storage overhead (that is, the RMS index tree
structure).
• Index records consist of block numbers, byte-in-block numbers and key.
• Can be write shared.
All VAX-11 RMS files have two additional blocks in the directory. These
additional blocks contain information relating the type of RMS file and the
record length.

D-4

Data Structures and Basic Concepts

Figure D.3: Sequential Files
END OF FILE~---

RECORD RECORD

RECORD

Q-MK-00027-00

Figure D.4: Relative Files
CELL NO.

999

1000

RECORD
999

I~.--------------------------BUCKET*------------------~--~-~1
.
*A bucket is a storage structure of 1 to 32 blocks.
Q-M K-00028-00

Figure D.5: Indexed Files

KEY
DEFINITION

~--PRIMARY INDEX (EMPLOYEE NAME)~

ABLE

ELM AV

24379] • • •

•• ••

JONES

•• ••

JONES

MAIN ST

19724

SMITH

HOLT RD

"------~---------------------DATA RECORDS----------------~----------

Q-M K-00029-00

Data Structures and Basic Concepts

D-5

D.2 Input and Output File Specification
An input or output file specification uniquely identifies a file by indicating its
physical location and a directory in which it is cataloged, as well as providing
a unique filename for that particular file within the directory. However, it is
not necessary to supply the physical location and directory for the file, since
the system uses the defaults set up during the log-in procedure when these
components are omitted from a file specification.
This section is only a summary. For a full description of defaults, wild cards,
logical names, and subdirectories, see the VAX/VMS Command Language
User's Guide.
The format of a file specification representing a physical file or device is:
node-naMe::deuice-naMe:[directoryJfilenaMe.file-tYPe;file-uersion

Node-name::

The individual computer system (or node) name within a network consists of 1-6 alphanumeric characters.
Example: BOSTON::

Device-name: The device name consists of three components: device type
[controller] [unit-number]:
The maximum length of the device type and controller specification is 15 characters. The maximum unit number is 65535.
The default value for controller is A, and the default value for
unit is 0.
Physical device names are:
Mnemonic
CR
DB
DM
DR

DX
LP

MB
MT
NET
TT

Device
Card Reader
RP04, RP05, RP06 Disk
RK06 Disk
RM03 Disk
Floppy Disk
Line Printer
Mailbox
TE16 Magnetic Tape
Network Communication Device
Interactive Terminal
DMC-11

Example: DB: is actually device name DBAO: by default.
[directory]

The directory name or names must be inclosed in either
square brackets ([]) or angle brackets ( <>). A directory without a directory name (for example, [ ]) is not valid. The directory types are:
• A 1- to 9-alphanumeric character string
• A two-part number in the format of a user identification
code (UIC)

D-6

Data Structures and Basic Concepts

• As subdirectories, in the format of name.name.name where
each name can consist of up to 9 alphanumberic characters;
each name represents a diretory level.

Filename.

The file name is limited to nine ASCII characters.

File type

The file type is limited to three ASCII characters. Some commonly used file types are:
File type
B2S
CMD
COM
COB
DAT*
DIF
DIR
DMP
EXE
FOR
LIB
LIS
LOG
LST
MAC
MAP
MAR
MLB
OBJ
ODL
OLB
OPT
STB
TSK

;File-version

Contents
Input source file for the PDP-11 BASIC-PLUS-2NAX compiler
Compatibility mode indirect command file
Command procedure file to be executed with the @ (Execute
Procedure) command, or to be submitted for batch execution
with the SUBMIT command
Input file containing source statements for the PDP-11
COBOL-74NAX compiler
Input or output data file
Output listing created by the DIFFERENCES command
Directory file
·
Output listing created by the DUMP command
Executable program image
Input file containing source statements for the VAX-11
FORTRAN-IV-PLUS compiler
Library file
Listing file created by a language compiler or assembler;
default input file type for PRINT and TYPE commands
Batch job output file
Compatibility mode listing file
MACR0-11 source file
Memory allocation map created by the linker
VAX -11 MACRO source file
Macro library
Object file created by a language compiler or assembler
Overlay description file
·
Object module library
Options file for input to the LINK command
Symbol table file created by the linker
Compatibility mode task image

The file version number is automatically updated by the system each time the file is changed. Commands may optionally
use a period to delimit the file version number, but the documentation will use a semicolon.

* indicates default file type for input files. Default file type for output files is whatever the
input file typu is.

Data Structures and Basic Concepts

D-7

D.3 Programming Languages Supported
The following compilers produce native mode programs that can use VAX-11
SORT:
• VAX-11 FORTRAN IV-PLUS
• VAX-11 MACRO
• VAX-11 BLISS
• VAX-11 COBOL-74
VAX-11 FORTRAN IV-PLUS

FORTRAN IV-PLUS is an especially complete version of the leading language for scientific and engineering computation. It is a high-performance
superset of the American National Standard lnstitute's (ANSI) 1966 FORTRAN. It also implements many of the anticipated features of the forthcoming ANSI standard.
FORTRAN IV-PLUS supports character data types, an IF-THEN-ELSE
statement, long variable names, and the standard CALL facility for calling
system services.
The FORTRAN IV-PLUS compiler first optimizes user source code, then
translates it to take advantage of the VAX-11 instruction set, which can
compile whole FORTRAN IV-PLUS statements into single instructions. An
interactive symbolic debugger allows source-level debugging of FORTRAN
IV-PLUS programs.
VAX-11 MACRO

The VAX-11 MACRO assembly language allows the programmer to write
32-bit machine language instructions for special efficiency. The symbolic debugger can also be used with VAX-11 MACRO.
VAX-11 BLISS

BLISS is a medium level language designed for building system software;
such as compilers, real-time processors, and utilities.
VAX-11 COBOL-74

The VAX-11 COBOL-74 language is based on the 1974 ANSI standard.

D-8

Data Structures and Basic Concepts

Index
Address files, 2:...8, 2-11
Address sort, 1-1, 2-8, 5-5
sample output, 5-6
selection of, 2-7
ASCII character set collating sequence, B-1
Batch sessions, 2-17
BLISS, D-8
Bucket, D-5
Buffer allocation and work areas, 5-10
Byte, C-1
Call by reference, 3-6
Calling SORT subroutines, 3-1
Character, D-1
Character string, C-5
COBOL-74, D-8
Command, 1-1, 2-2
description, 2-17
summary, 2-3
Command description,
command name qualifiers, 2-18
input-file-specification qualifiers, 2-20
output-file-specification qualifiers, 2-21
Command format,
continuous command string, 2-1
using system prompts, 2-1
Command qualifier,
/KEY, 2-18
/PROCESS, 2-17
/RSXll, 2-19
/SPECIFICATION, 2-19
/WORK_FILES, 2-19
Command structure,
command name, 2-1
command parameters, 2-2
command qualifiers, 2-2
file qualifiers, 2-2
input-file-specification, 2-2
keywords, 2-2
output-file-specification, 2-2
subqualifiers, 2-2
values, 2-2
Command summary, 2-3
CPU time, 5-14
Data files, 2-11
input, 5-3
output, 5-4
Data hierarchy, D-1

Data types, C-1
byte, C-1
character string, C-5
double floating, C-3
floating, C-3
leading separate numeric string, C-7
longword, C-2
packed decimal string, C-8
quadword, C-2
trailing numeric string, C-5
variable length bit field, C-4
word, C-1
Decimal to hexadecimal conversion A-2
Decimal/Octal conversion, A-1
'
Descriptor, 3-5
Device name, D-6
DIGITAL command language (DCL) 1-1
Directory, D-6
'
Double floating datum, C-3
Dynamic memory usage, 5-10
Efficiency, 5-1
Elapsed time, 5-14
Error conditions,
fatal, 4-1
warning, 4-1
Error messages, 4-1
SORT, 4-2
VAX-11 RMS, 4-8
VAXNMS DCL command interpreter, 4-2
Fields, D-1
File,
I/0 considerations, 2-9
1/0 interface, 3-1
name, D-7
size, 3-6
type, D-7
version, D-7
File organizations,
1/0 flow, 2-10
indexed, D-3
indexed-sequential data files, 2-11
relative, D-3
relative data files, 2-11
Sequential, D-3
sequential address file, 2-11
sequential data files, 2-11
File specification format, D-6

Index-1

Files, D-1
Fixed length record sample, D-2
Floating datum, C-3
FORTRAN IV PLUS, D-8
Functionality, 1-2
Hexadecimal integer columns, A-3
Hexadecimal to decimal conversion, A-2

I/0 considerations, 2-8, 2-9, 5-11
1/0 counts, 5-14
Index sort, 1-1, 2-8, 5-6
sample output, 5-7
selection of, 2-7
Indexed files sample, D-5
Indexed-sequential data files, 2-11
Input and output, 1-2
Input and output file specificatiori, D.:...6
Input buffer, 5-10
Input data files, 5-3
Input file,
descriptor, 3-9
Input file format specifications,
file size, 2-21
record size, 2-20
Input file qualifiers,
/FORMAT, 2-20
Input record,
sample, 5-4
types, 5-4
Inputs to VAX-11 SORT, 2-·8, 2-9
Interactive mode, 2-1
Interactive samples, 2-12
Interactive sessions, 2-5
sample -#1, 2-13
sample -#2, 2-15
Invoking SORT, 2-5
Key, 2-31
Key area, 3-5
Key buffer,
address, 3-6
specifications, 3-6
Key comparison routine address, 3-6, 3-7
Key comparisons,
SORT's routine, 3-4
user's routine, 3-4
Key specifications,
ascending/descending, 2-19
data type, 2-18
leading/trailing sign, 2-19
number, 2-18
overpunched/separate sign, 2-19
position, 2-18
size, 2-18

Index-2

Keys,
ascending or descending order, 2-35
assigning a precedence number, 2-31
binary, 2-33
character, 2-33
data type, 2-32
decimal, 2-33
multiple, 2-34
NUMBER=n, 2-31
POSITION=n, 2-34
quick reference flowchart, 2-31
setting up, 2-31
signed numbers, 2-33
size restrictions, 2-35
SIZE=n, 2-34
Keywords, 2-2
Languages, D-8
Leading separate numeric string, C-7
Longest record length, 3-6, 5-13
Longword, C-2
MACRO, D-8
Mapping, 5-10
Multi-block and buffer counts, 5-14
Node name, D-6
Number of,
work files, 3-6
initial runs, 5-14
merge passes, 5-14
Octal/Decimal conversion, A-1
Order of the merge, 5-14
Output buffer, 5-10
Output data files, 5-4
sample, 5-5
Output file,
allocation, 3-9
bucket size, 3-9
descriptor, 3-9
file options, 3-9
maximum record size, 3-9
organization, 3-9
record format, 3-9
Output file format specifications,
block size, 2-22
record size, 2-21
record type, 2-21
Output file qualifiers,
/ALLOCATION, 2-22
/BUCKET_SIZE, 2-22
/CONTIGUOUS, 2-22
/FORMAT, 2-21
/INDEXED_SEQUENTIAL, 2-22

Output file qualifiers (Cont.)
/OVERLAY, 2-22
/RELATIVE, 2-22
/SEQUENTIAL, 2-22
Output record placement, 3-13
Outputs from VAX-11 SORT, 2-9
Overpunch format, C-6
Packed decimal string, C-8
Page faults, 5-10, 5-14
Paged-out, 5-10
Pager, 5-10
Parameters, 2-2
Phases of operation, 5-7
Powers of 2 and 16, A-2
Programming languages supported,
BLISS, D-8
COBOL-74, D-8
FORTRAN IV PLUS, D-8
MACRO, D-8
Quadword, C-2
Qualifiers, 2-2, 2-3
Real memory, 5-10
Record descriptor, 3-11, 3-13
Record I/0 interface, 3-2
Record sort, 1-1, 2-8, 5-3
selection of, 2-7
Record types,
Fixed, D-2
Variable, D-2
Variable with fixed-length control (VFC), D-2
Records, D-1
Re la ti ve data files, 2-11
Relative files sample, D-5
RMS,
completion status codes, 4-8
data files, D-1
data structures, D-1
file organizations, D-3
records, D-1
Running SORT, 2-1
Sample programs,
COBOL-74NAX, 3-17
FORTRAN IV PLUS, 3-20
MACRO, 3-15
Sequential,
address files, 2-11
data files, 2-11
files sample, D-5
SORT,
architecture, 5-2
definition of, 1-1

SORT (Cont.)
functional description, 5-1
internal organization, 5-7
Sort,
improving efficiency, 5-1
sample, 2-6
tree, 5-10
tuning, 5-1
type, 3-6
types, 1-1
Sort types,
address, 1-1, 2-7, 5-6
index, 1-1, 2-7,5-6
record, 1-1, 2-7, 5-3
selection of, 2-7
specification of, 2-17
tag, 1-1, 2-7, 5-3
SORT's four sorting processes, 2-8
Sorting processes, 2-8
functional descriptions, 5-3
See sort types, 1-1
Sorting sequence, 2-5
Specification file, 2-23
fixed position field,
format (SORT-11), 2-24
field record specfications, 2-27
header record specifications, 2-26
sample, 2-29
summary, 2-28
free field,
format (VAX-11 SORT), 2-24
field record specifications, 2-30
header record specifications, 2-30
sample, 2-31
methods of entering, 2-23
record formats for fixed position fields, 2-25
record formats for SORT-11 type files, 2-25
records, 2-24
VAX-11 SORT, 2-30
Specification form, 2-25
Statistics, 1-2
example of, 2-12
String descriptors, 3-6
Subqualifiers, 2-2, 2-3
Subroutines, 3-1
calls, 3-1
interfaces, 3-1
package, 1-1
parameters, definitions, and valid returns, 3-5
programming considerations, 3-2
set summary, 3-3
SOR$END_SORT, 3-14
SOR$INIT_SORT, 3-6
SOR$PASS_FILES, 3-9

Index-3

Subroutines (Cont.)
SOR$RELEASE-REC, 3-11
SOR$RETURN-REC, 3-13
SOR$SORT_MERGE, 3-12
Tag sort, 1-1, 2-8, 5-3
selection of, 2-7
Temporary storage, 2-8
Total key size, 3-6
Trailing numeric string, C-5
Tuning procedure, 5-11
system manager performance
considerations, 5-14
user performance considerations, 5-11
Tuning procedure parameters,
modified page writer clustel' factor, 5-15
number of work files, 5-13
process section count, 5-15
type of sort, 5-13
using SORT'S statistics, 5-13
virtual page count, 5-15
work file devices, 5-13
working set quota, 5-12, 5-15

Index-4

Utility, 1-1
Values, 2-2, 2-3
Variable length bit field, C-4
Variable length record samples, D-2
VAX-11 SORT,
definition of, 1-1
VAXNMS calling standards, 3-2
VFC records, D-2
sample, D-2
Virtual memory, 5-10
Virtual memory added, 5-14
VMS library, 3-2
Word, C-1
Work files, 2-8
default, 2-35
logical name of, 2-35
physical device codes, 2-35
setting up, 2-35
Working set, 5-10
Zoned numeric format, C-6

READER'S COMMENTS

VAX-11
SORT
User's Gulde
Order No. AA-D113A-TE

NOTE: This form is for document comments only. DIGITAL will use comments submitted on this form at the
company's discretion. If you require a written reply and are eligible to receive one under Software
Performance Report (SPR) service, submit your comments on an SPR form.

Did you find this manual understandable, usable, and well-organized? Please make suggestions for improvement.

Did you find errors in this manual? If so, specify the error and the page number.

Please indicate the type of user/reader that you most nearly represent.
D
D
D
D

D
D

Name

Assembly language programmer
Higher-level language programmer
Occasional programmer (experienced)
User with little programming experience
Student programmer
Other (please s p e c i f y ) - - - - - - - - - - - - - - - - - - - - - - - - - - -

Date-----------------

Organization---------------~------------------

Street------------------------------------C i t Y - - - - - - - - - - - - - - - - - - State------ Zip Code _ _ _ _ _ _ __
or
Country

Do Not Tear - Fold Here and Tape -

111111

POSTAGE WILL BE PAID BY ADDRESSEE

RT/C SOFTWARE PUBLICATIONS TW/A14
DIGITAL EQUIPMENT CORPORATION
1925 ANDOVER STREET
TEWKSBURY, MASSACHUSETTS

Do Not Tear- Fold Here

01876

No Postage
Necessary
if Mai led in thE
United States