Digital PDFs

AA-3341C-TC

May 1977

113 pages

Original

3.9MB

Document:	PDP-11 SORT Reference Manual
Order Number:	AA-3341C-TC
Revision:	000
Pages:	113
Original Filename:

OCR Text

PDP-11
SORT Reference Manual
Order No. AA-3341C-TC

May 1977
This document discusses the features and uses of the PDP-11 SORT utility.
Installation procedures for the operating systems listed below are included.
This document and the software to which it pertains are for use with
RMS-11-formatted tiles only.

PDP-11
SORT Reference Manual
Order No. AA-3341C-TC

SUPERSESSION/UPDATE INFORMATION:

This document supersedes the document of the same name,
Order No. DEC-11-USTMA-B-D, published July 1975 for
the operating systems listed below only.

OPERATING SYSTEM AND VERSION:

RSX-11 M V03 and V3.1
RSTS/E V06B
IAS V02

SOFTWARE VERSION:

SORT-11 V02

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, October 1974
Revised:
July 1975
May 1977

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 software described in this document is furnished under a license
and may be used or copied only in accordance with the terms of such
license.
Digital Equipment Corporation assumes no responsibility for the use
or reliability of its software on equipment that is not supplied by
DIGITAL.

The postage prepaid REl\DER'S COMMENTS form on the last page of this
document requests the user's critical evaluation to assist us in preparing future documentation.
The following are trademarks of Digital Equipment Corporation:

DIGITAL
DEC
PDP
DEC US
UNIBUS
COMPUTER LABS
CO MT EX
DDT
DEC COMM

DECsystem-10
DECtape
DIBOL
EDUSYSTEM
FLIP CHIP
FOCAL
INDAC
LAB-8
DECsystem-20

MASSBUS
OMNIBUS
OS/8
PHA
RSTS
RSX
TYPESET-8
TYPESET-10
TYPESET-11

CONTENTS

Page
vii

PREFACE
CHAPTFR

CHAPTER

OVERVIEW

1-1

1.1
1.1.1
1.1. 2
1. 2
1.3
1. 3 .1
1. 3. 2
1. 3. 3
1. 3. 4
1. 3. 5

1-1
1-2
1-4
1-7
1-8
1-8
1-9
1-9
1-10

INTRODUCTION
Data Files
Command String and Specification File
HOW SORT OPERATES
SORT PROCESSING OPTIONS
Record Sort (SORTR)
Tage Sort (SORTT)
Address Routing Sort (SORTA)
Index Sort (SORTI)
Graphic Representation of SORTA and SORTI
output
THE SORT COMMAND STRING

2.1
2.2
2.2.1
2.2.2
2.3
2.3.1

INTRODUCTION
HOW TO RUN SORT
Running SORT in Interactive Mode
Running SORT in RSTS/E Batch Mode
THE COMMAND STRING FORMAT
File Specifications

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

CONTROLLING SORT OPTIONS

3-1

3.1
SWITCHES
3 .1.1
ALLOCATION (/AL:n)
3 .1. 2
BLOCKSIZE (/BL:n)
3 .1. 3
BUCKETSIZE (/BU:n)
3 .1. 4
CONTIGUOUS (/CO)
3 .1. 5
DEVICE (/DE:x)
FILES (/FI:n)
3 .1. 6
3 .1. 7
FORMAT (/FO:x:n)
3 .1. 8
INDEXED SEQUENTIAL (/IN:X)
KEY (/KE:abm.n)
3 .1. 9
3.1.10
PROCESS (/PR:x)
RELATIVE (/RE)
3.1.11
SEQUENTIAL (/SE)
3 .1.12
SIZE (/SI:n)
3.1.13
3.2
THE SPECIFICATION FILE
3.2.1
Header Specification
3.2.1.1
Notes and Comments on Header Specification
Entries
ALTSEQ Format and Notes
3.2.1.2
3.2.2
Record Type Specification
3.2.2.1
Notes and Comments on Record Type
Specifications
3.2.3
Field Specifications
Notes and Comments on Field Specification
3.2.3.1
Entries
iii

1-10
2-1

3-1
3-1
3-1
3-1
3-3
3-3
3-3
3-3
3-3
3-4
3-5
3-5
3-5
3-5
3-5
3-9
3-10
3-10
3-11
3-13
3-14
3-16

CONTENTS (Cont.)

Page

CHAPTER

3.2.4
3.3

Using a Specification File
SORT SPECIFICATION SUMMARY

3-17
3-19

ERROR CONDITIONS

4-1

4.1
4.2
4.3
4.4
4.5
4.5.1
4.5.2
4.6
4.7

ERRORS IN SORT: GENERAL COMMENTS
COMMAND DECODER ERRORS
SPECIFICATION FILE ERRORS
I/O ERRORS
CONTROL PROGRAM ERRORS
Pre-Sort Errors
Merge Errors
OTHER ERRORS
RMS ERROR CODES

4-1
4-1
4-3
4-5
4-6
4-6
4-6
4-6
4-7

INTERNAL OPERATION

5-1

5.1
FILES
5 .1.1
User and Scratch File Name Conventions
5.1. 2
LUN Assignments
5.1. 3
File Maintenance
5 .1. 4
File Allocation
5.1. 5
Scratch File Structure
5.1. 6
Scratch Files Use
5.2
ORGANIZATION OF SORT
SORTS ERRORS
5.3
5.4
USING THE SORTS SUBROUTINE PACKAGE
5.4.1
Conventions and Standards
RSORT - Initializing the Sort
5.4.1.l
5.4.1.2
RELES - Passing Input Records to the Sort
5.4.1.3
MERGE - Merging the Scratch Files
5.4.1.4
RETRN - Requesting an Output Record
ENDS - Ending the Sort
5.4.1.5
Setting up the Keys
5.4.2
5.4.3
Calculating the Size of the Work Area
5.4.4
Typical Calling Sequences
5.4.5
Linking SORTS Subroutines with Your Program

5-1
5-1
5-1
5-2
5-2
5-3
5-3
5-4
5-4
5-5
5-6
5-6
5-7
5-8
5-8
5-9
5-9
5-10
5-10
5-10

APPENDIX A

CHARACTER SETS USED BY SORT

A-1

APPENDIX B

P:RINTABLE CHARACTERS

B-1

APPENDIX c

SORT PROGRAM EXAMPLES

C-1

APPENDIX D

INTERNAL OPERATIONS

D-1

PROCESS DESCRIPTION
THE SORT WORK AREA
PERFORMANCE PARAMETERS

D-1
D-1
D-4

SORT INSTALLATION

E-1

GENERAL COMMENTS
SORT KIT CONTENTS
Building the SORT Task for RSX-llM or IAS
Building the SORT Task for RSTS/E
CHANGING THE SORT TASK IMAGE SIZE

E-1
E-1
E-2
E-2
E-3

D.l
D.2
D.3
APPENDIX E
E.l
E.2
E.2.1
E.2.2
E.3

CONTENTS (Cont.)

Page
APPENDIX F

F-1

RMS I/O STATUS CODES

Glossary-1

GLOSSARY
INDEX

Index-1

FIGURES
FIGURE

1-1
3-1
3-2
3-3
C-1
C-2

C-3
C-4
c-5a

C-5b
D-1

Sample Record Types
Specification File Format
SORT Specification Form
SORT Specification Entries
Sales List
Overtime Analysis
Employee List
Bonus Analysis
Stock Order List, Page 1
Stock Order List, Page 2
SORT Work Area Allocation

1-3
3-6
3-8
3-18
C-2
C-3

c-4

C-5
C-6
C-7
D-3

TABLES
TABLE

1-1
1-2
1-3
2-1
3-1
3-2
3-3
3-4
4-1
4-2
4-3
A-1

A-2
B-1

B-2
D-1
F-1

Selecting the Sorting Process and Devices
that Best Suit the Processing Environment
Sorted Output File
Sorting Process Options
Device Specification Codes
Switch Summary
Header Specifications
Record Type Specifications
Field Specifications
RMS Status Codes for RSX-llM
RMS File Attribute Mismatch Status Codes
Other Commonly Occurring Status Codes
The ASCII Character Set with Corresponding
EBCDIC Codes
The Subset of the EBCDIC Character Set used
by SORT when EBCDIC Sorting is Requested
Printable Characters Grouped by Equal Digits
Printable Characters Grouped by Equal Zones
User-Accessible SORT Parameters
RMS Error Status Codes

1-3
1-4
1-9
2-4
3-2
3-19
3-20
3-22
4-7
4-7
4-8

A-2
A-3

B-2
B-3
D-5

F-1

PREFACE

This reference manual describes the features and operation of the SORT
utility program. Chapter 1 describes the SORT program, its operation
and environment. Chapter 2 is an operator's guide for running SORT.
Chapter 3 is a reference for those who write the instructions for the
SORT program. Chapter 4 describes error conditions, and Chapter 5
contains more detailed information about the SORT program for special
applications.
Appendices A and B contain tables of ASCII and EBCDIC, and zone/digit
card punch codes respectively. Appendix C contains several examples
of SORT programs. Appendix D discusses internal SORT operations,
mainly for the benefit of experienced programmers.
Appendix E
Appendix F
contains SORT installation information for all systems.
contains RMS-returned status codes.
A glossary is included to define terminology used in this manual.

vii

CHAPTER 1
OVERVIEW

1.1

INTRODUCTION

The SORT utility program allows you to read any input file, to sort
the contents, and to write out the sorted data onto an output file.
The sorting sequence is determined by control fields, also known as
key fields, within the data itself. If you do not wish to sort your
data base, you can still use SORT to extract key information, sort
that information, and store the sorted information on a permanent
file. You can later use that file to access your data base in the
order of the key information on the sorted file. The contents of the
sorted file may be entire records, key fields, or record indices
relative to the position of each record within the file (the first
record on the data base is record 1, the second, 2, etc.).
SORT provides four sorting techniques:
•

Record Sort (SORTR) produces a reordered file by using the
entire contents of each record as the record key. A record
sort can be used on any acceptable input device and can
process any valid RMS (Record Management Services) format.

•

Tag Sort (SORTT) produces a reordered file by sorting only the
record keys. SORT then randomly reaccesses the input file to
create a resequenced output file according to those record
keys.
The tag sort method conserves temporary storage, but
can only accept input files residing on disk.

•

Address Routing Sort (SORTA) produces an address file without
reordering the input file. The address file, sorted by record
keys, can later be used as an index file* to read the data
base 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 customer-number index or
sales-territory index for different reports.
SORTA is the
fastest of the four sorting methods.

•

Index Sort (SORT!) produces an index file* containing the key
field of each data record and a pointer to its location in the
input file. The index file can be used for sequential or
direct accessing from a random file.

The SORT utility program may be controlled by a command string and an
optional specification file. There is a simple format for each. If
your SORT application does not require that records be restructured or
that only a subset of the input file be sorted, then you need only a
command string to control SORT.
*

Not indexed by RMS-11.
1-1

OVERVIEW

SORT can handle any RMS-valid file organization.
Different file
organizations are distinguished by the ordering of the records they
contain and the way they handle the retrieval process.
The order of the records in a sequential file is determined by the
order in which the records are read from the file. The first record
in the file is the first record read out, regardless of whether the
records are written to the file in some sorted order or not.
A relative file consists of record areas that are identified by
relative record numbers. The first record area in the file is record
number 1, the second is 2, etc., much the same as an apartment house
where the first apartment is 1, and so forth. But, as in an apartment
house, if you want the record that is in the twelfth record area, for
instance, you must ask for record number 12, even though there may not
be records in areas 1 through 11. Relative files can reside only on
disk.
An indexed file contains prologue information, one or more indices,
and the data records themselves. 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.
Table 1-1 shows the devices that you may use to supply data
Data may be stored in binary, ASCII, or EBCDIC form.

1.1.1

SORT.

Data Files

SORT may accept a file from any one
available in the system configuration:
1.

Disk unit

Magtape unit

Card reader

Paper tape reader

Terminal

the

peripheral

devices

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 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 1-1 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 8 to 13. The "total amount of sale" is an
example of a data field not relevant to the sorting process.

1-2

OVERVIEW
Table 1-1
Selecting the Sorting Process and Devices
that Best Suit the Processing Environment

---i ------ -·-·-

Work
File

Output
File

I np ut
F il e

Sorting
Technique

·--

- - - ---

···-

·--

Disk
(3-8 files)

SORTR

Dis k

Disk

(Record Sort)

Mag ta pe

Magtape 1

Pap er Tape

Paper Tape

Car ds

Printer
Console

Con so le
-

- · -t - · - -

Disk
(3-8 files)

SORTT

Dis k

Disk

(Tag Sort)

Mag ta pe

Magtape 1
Printer
Console
Paper Tape

-----------'

Disk
(3-8 files)

Disk

SORTA

Dis k

(Address
Routing Sort)

Mag ta pe

SORT I

Dis k

+----

Disk
(3-8 files)

Disk

Mag ta pe

(Index Sort)
L-------·--··----

'---·~---·

~-··-·

----·

-----~-·--·

-·~·-

--~~~

1 Provided

records are at least 18 bytes long.
be in ANSI format.

Magtape must

RECORD
IDENTIFIER

SALES
RECORD

ORDER
RECORD

DESCRIPTION
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

Figure 1-1

Sample Record Types

1-3

OVERVIEW

If you request a sort in ascending order on the sales records
illustrated above, 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

Number of items sold

The output file contains all sales records in the order illustrated in
Table 1-2. Please note that data fields other than those you use for
keys, salesperson's code for instance, may be in any order.
Table 1-2
Sorted Output File
Major Key: Item Number

Minor Key: Quantity

Lowest item no.

Lowest quantity
Next higher quantity

Lowest item no.

Highest quantity

Next higher item no.

Lowest quantity
Next higher quantity

Next higher item no.

Highest quantity

Highest item no.

Lowest quantity
Next higher quantity

Highest item no.

Highest quantity
.. ----·----·--·--·--- ·-······-------·

1.1.2

Command String and Specification File

You can direct the SORT program by entering
command string has three functions:
1.

To reference devices in the
current sort
1-4

system

command
for

each

string.

The

file

the

OVERVIEW
2.

To specify switches that define file parameters used
sorting process

To reference a specification
switches to control the sort

file

specify

the
other

Several command string switches define the sorting process parameters.
One switch describes record formats and the maximum record size.
Another delimits the internal work files.
Others provide detailed
file information to RMS.
Normally, you direct the sort with a specification file.
Two
additional switches may be used instead of a specification file to
control a sort. One switch specifies the sorting process option; the
other identifies the key fields. The use of these switches is limited
to sorting an input file of uniform format:
1.

The key fields must reside in
record of the iriput file.

the

same

The file must contain only the records to be included in the
sort. The figure below illustrates a general sort that would
only require a command string and switches.

l c

location

every

SORT USING COMMAND STRING AND SWITCHES

The specification file is the supplement to the command string, which
provides the basis for controlling and directing the sorting process.
The specification file gives you a variety
They are listed below:
1.

controlling

Record Selection

RECORD SELECTION

1-5

features.

OVERVIEW
You can include or omit any records from the sorting process.
The output file will contain only the records that you
specify.
2.

Alternate Collating Sequence
":m::H•m•

B
A

c
D
A

- - + e-o.o-.a .._.

t....J

I-

I-'

--ALTERNATE COLLATING SEQUENCE
11-1523

If necessary, you can specify an
alternate
collating
sequence. The normal sequence is that implied in ASCII code.
One alternate choice is EBCDIC values.
The other is an
individual alternate collating sequence (ALTSEQ) . An ALTSEQ
can be used to change the ASCII values of the normal
sequence. It applies to all the alphanumeric key data in the
records, but only during the actual sorting process.
The
output record remains unchanged.

Forced Keys

~:
102

351
242

-.-...3nn-/nn--...

242
102
351

333

I-

II-'

t....J

I-'

___..

-··-·
FORCED KEY
11-1524

An ALTSEQ applies to all positions of the key.
Forced keys
allow you to specify an alternate sequence for particular
positions within the key.
You can specify an alternate
sequence by substituting a lower-valued character, such as
the slash (/) in the example above. Since the slash comes
before O, the 300-series records in the example are brought
to the front of the file. Notice that the records so treated
are in sequence and in front of the rest of the sorted file.
The net effect is that of two sorted files, one behind the
other.

1-6

OVERVIEW

Input Format Variation

I . .,
l

l
J

N 141

Abbby

A bbby

A oaoz

N 207

N 141

INPUT FORMAT VARIATION

If the input file contains records with several different
formats, you can identify those records by type so that they
may be properly handled.
In the example above, A and N are record identifiers.

Output Format Variation

B mn 19

D UV 23

C SI 17

D 23

SIC 17

mn B 19
xy A 11

A 'Y 11

I-'

IfOUTPUT FORMAT VARIATION

You can change the format of the data file during the sort,
but you cannot change the contents of any given data item.

1.2

HOW SORT OPERATES

The SORT program consists of two basic parts: a control program and a
subroutine package called SORTS.
The control program directs the
overall processing. SORTS serves as a collection of subroutines that
the control program uses during its processing. You can write your
own control program to take advantage of the
separate
SORTS
subroutines.
(See Section 5.4.)
There are three phases of operation in the SORT control program.* In
the first phase, SORT reads the command string, decodes, and stores

* Each phase of operation corresponds
program structure.
1-7

overlay

the

SORT

OVERVIEW

the switch values and the specification file, if present. Any errors
in the command string or specification file are reported at this
point.
The second phase begins the pre-sort operation. The control program
is called to open and read the input file and establish the keys.
Then the SORTS subroutine begins 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 scratch files on bulk storage devices.
In order to merge these files and complete the sort, space for at
least three scratch files must be available.
The SORT program
normally provides for a maximum of eight scratch files. Either a
switch in the command string or the amount of available internal work
space can reduce the number of scratch files used.
The merge phase rebuilds the intermediate scratch files into a merged
file.
Another subroutine reads the records in the proper sequence.
The records are then written into the output file. If there are no
scratch 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 scratch files and returns to the first phase; SORT is then ready
to accept another job.

1.3

SORT PROCESSING OPTIONS

There are four SORT processes: Record Sort, Tag Sort, Address Routing
Sort, and Index Sort.
You can specify the SORT process in the
specification file or with a switch in the command string.
The
default process is the Record Sort. Each process has its particular
input requirements, processing methods, and resultant output files.
In general, Record Sort performs a relatively slow sort.
Tag Sort
produces the same kind of output as Record Sort. Tag Sort performs a
faster sort if the key size is much smaller than the total record
size.
The Address Routing and Index Sorts are high-speed sorts that
produce compact output. See Table 1-3.
1.3.1

Record Sort (SORTR)

The Record Sort (SORTR) outputs all specified record data in a sorted
sequence.
Each record is kept intact throughout the entire sorting
process. Since it moves the whole record, SORTR is relatively slow
and may require considerable main memory or external storage work
space for large files.
NOTE
Records can be FIXED, STREAM ASCII, or
VARIABLE length in any valid RMS (Record
Management Services) format.
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.
1-8

OVERVIEW
Table 1-3
Sorting Process Options
.-----------·--

Type of
SORT

Type of
File

---

Record Size
and Format

Device

Speed

•-.----- -- ----- - - - - - - - - - - - - - -

SORTR
(Record
Sort)

Input
and
Output

Any

SORTT
(Tag
Sort)

Input

Any

Any appropriate
device including:
disk, rnagtape 1
card reader,
console
"·--·

- --··

-------------t-- - ---------- -------- ---··---·----

Disk

-----------

Output

Any

Input

Any

Slowest

Any appropriate
device (including
rnagtape 1 )

Slow for
large files,
large keys

~--··-----

SORTA
(ADDROUT
Sort)

Disk or rnagtape 1

Output

Fixed, six
bytes

Input

Any

-Disk

~-----

SORT!
(Index
Sort)

-+-----------·--+--------------1
1

Disk or rnagtape

Fast

----------------!

Output

Fixed, 6byte pointer
+ original
key

Magtape labels are ANSI standard.
tape labels.

1.3.2

Fastest

- - - - - - - - - - · - - - - - - -1

Disk

DOS is not supported for rnag-

Tag Sort (SORTT)

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

1.3.3

Address Routing Sort (SORTA)

SORTA produces address routing files, which consist of relative record
pointers, beginning at 1, 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 SORTT, SORTA uses the minimum amount of
data necessary in the sorting process.
Once the input phase is
completed, the input file is not read again.
The output data is in a
restricted mode.
This means that SORTA is the fastest sorting method
in the sort package.
Do not transfer an ADDROUT index file
to a
device that cannot handle binary data, such as a printer.

1-9

OVERVIEW

1.3.4

Index Sort (SORT!)

SORTI produces an index file (not indexed by RMS-11) consisting of
relative record pointers, as in SORTA, and index keys. This makes it
slightly slower than SORTA. During processing SORT! handles only the
relative record pointers 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.
NOTE
The SORT! output file contains two extra
records at the beginning of the file:
first record - output record
count +2
second record - input record
count
Each
of
these
additional
records
contains
a key consisting of ASCII
nulls.
(See Table 1-3.)
If the key size is odd, the key portion of the output record
padded with a null byte.

1.3.5

will

Graphic Rep.resentation of SORTA and SORT! Output

The ADDress ROUTing Sort (SORTA) produces an output file consisting of
record indices.
Each record index occupies one 6-byte record in the
output file. Here is an example of SORTA output:
Let us assume that we are sorting a file consisting of six
records
with
SORTA.
If the sequence of record indices
corresponding to the sorted records is 5,1,6,3,4,2 then the
output from SORTA can be represented as follows:

1-10

OVERVIEW

BLOCK NUMBER
HIGH
LOW
(2 bytes
(2 bytes
16 bits)
16 bits)

RECORD
NUMBER

BYTE-I N-BLOCK
(2 by tes
16 b its)
-

~--

0 0 0 1 6 2

0 0 0 0 0 1

0 0 0 0 0 0

0 0 0 0 0 1

0 0 0 0 0 0

0 0 0 0 0 1

0 0 0 0 0 0

0 0 0 2 3 6

0 0 0 0 0 1

0 0 0 0 0 0

0 0 0 0 4 2

0 0 0 0 0 1

0 0 0 0

0 0 0 1 3 2

0 0 0 0 0 1

0 0 0 0 0 0

0 0 0 0 2 6
----

--------

Index Sort (SORT!) produces 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 in original form.
Following is an example of SORT! output.
Assume that you are sorting a file consisting of six records with
SORT!, and you are using a key size of four characters (bytes). Note
that SORT! writes two additional records at the beginning of the
output file:
the first is the output record count plus 2; the second
is the input record count.
The sequence
of
record
indices
corresponding to the sorted record, is 5,1,6,3,4,2. The increased
output record count in the first additional record reflects the fact
that SORT! added two records to the file. Here is part of block 1:
~-

RECORD
NUMBER

BLOCK NUMBER
LOW
HIGH
( 2 bytes
( 2 bytes
16 bits)
16 bits)
----------

"""---.------

BYTE-IN-BLOCK
(2 bytes
16 bits)
-1-"

··--·--

---·----

----

KEY IN ORIGINAL FORM

1--" ----

---·-·-~

0 0 0 0 0 1

0 0 0 0 0 0

0 0 0 0 1 0

0 0 0 0 0 0

0 0 0 0 0 1

0 0 0 0 0 0

0 0 0 0 0 6

0 0 0 0 0 0

0 0 0 0 0 1

0 0 0 0 0 0

0 0 0 1 6 2

0 0 0 0 0 1

0 0 0 0 0 0

0 0 0 0 0 1

0 0 0 0 0 0

0 0 0 2 3 6

0 0 0 0 0 1

0 0 0 0 0 0

0 0 0 0 4 2

0 0 0 0 0 1

0 0 0 0 0 0

0 0 0 1 3 2

0 0 0 0 0 1

0 0 0 0 0 0

0 0 0 0 2 6

NOTE
ABCD represents the sorting keys in
original format.
The record count in
the first record is an octal 10 (decimal
8) •

1-11

CHAPTER 2
THE SORT COMMAND STRING

2.1

INTRODUCTION

You initiate the SORT utility program by executing a SORT command
string.
This chapter describes the command string format for RSTS/E
and RSX-llM users. The IAS command string is described in the IAS
User's Guide.

2.2

2.2.1

HOW TO RUN SORT

Running SORT in Interactive Mode

To invoke SORT use one of the
prompt:

following

commands

after

the

system

SRT
RUN [1,2] SORT
RUN$SORT
SORT then prints SRT> to indicate that it is ready to accept a command
string.
You must specify one input file and one output file. A
second input file, if you give one, is the specification file.
One level of indirect command files is
command string is shown below:

permitted.

example

SORT

SRT>OUTPUT=INPUT/FO:VAR:lOO,SPEC
This command string directs SORT to sort the file INPUT.DAT on the
system device and place the sorted file OUTPUT.DAT on the system
device. The specification file SPEC.SRT on the system device is used
to define the sorting parameters. The switch /FO:VAR:lOO is described
in Section 3.1.7.
After you enter the SORT command string, SORT begins its processing.
If there are errors in the command string or specification file, they
are immediately reported on the console. At the end of a successful
run, SORT prints the following message:
SRT
M:ELAPSED REAL TIME: hh:mm:ss
SRT -- M:TOTAL RECORDS SORTED: nn
This message tells you the real or wall clock time necessary
perform the sort and the number of records in the output file.

2-1

THE SORT COMMAND STRING

When the sort is done, SORT again prompts with SRT>.
At this point
you can either enter another command string or terminate execution
with AZ (CTRL/Z) (or Ac - RSTS/E only).

2.2.2

Running SORT in RSTS/E Batch Mode

To run SORT in RSTS/E batch mode, precede your command string with the
following:
$RUN $SORT
$DATA
The command strinq is the same as
section.

2.3

that

described

the

preceding

THE COMMAND STRING FORMAT

The SORT program requires a command string to direct each job.
The
command string names the input and output files and the devices on
which they reside. Command string switches define the input and
output file parameters. The specification file or additional command
string switches control the sort. The use of the switches and the
specification file is explained in Sections 3.1 through 3.3.
A command string has one of the following forms:
output-files-list=input-files-list
@command-file
Both output-files--list and input-files-list are strings consisting of
file specifications separated by commas.
Each file specification
selects a file to be used as an output or input file by the system
program.
In particular, the SORT output-files-list consists of one file
specification;
and the SORT input-files-list consists of one or two
file specifications. The second input file is optional, and is
described in Section 3.2 as the specification file.
A file specification may have one or more switches associated with it.
A switch is an indicator used to select program options. Switches for
SORT have the following form:
/name:value
The switch name is a string of characters of which
two are significant.

only

the

initial

The second form of command string specifies an indirect command file.
Each line on the file is used as a command string (as if you had typed
it at the terminal) before any command strings following the file
specification are accepted.
The command string syntax is described more completely in the
Programmer's Reference Manual for your system or, for RSTS/E users,
the RSTS/E User's Guide.

2--2

THE SORT COMMAND STRING

2.3.l

File Specifications

A file specification has the following form:
dev: [uic]filenam.typ;ver/swl:vl: ..• :vn/sw2:vl ...
where:
dev:

2-character alphabetic legal device code
optional 1- or 2-digit unit number.
default is SY, the system disk.

and
The

[uic]

project-programmer number
specifying
the
directory.
The UIC is optional, but, if
used, must be a valid User Identification
Code.

filename

any 1- to 6-character (RSTS/E only) or 1- to
9-character
alphanumeric
string
that
specifies a file name.

.typ

1- to 3-character file type.
System programs
default the file type to an appropriate
standard type. So, the typical user will not
need to specify the file type.

;ver

version number of the file (for RSX-llM and
IAS users only -- omit for RSTS/E). This can
normally be omitted because
the
system
defaults to the highest existing version
number on input and to the highest existing
version number plus 1 on output.

File types default as shown below:
File

Default Type

Input

DAT

Output

DAT

Specification

Command

CMD

File specifications are discussed in detail in the Programmer's
Reference Manual or, for RSTS/E users, the RSTS/E User's Guide.

2-3

THE SORT COMMAND STRING
Table 2-1
Device Specification Codes

Dev:

Device*

SY:

System disk, default device

DKn:

Disk, cartridge unit n; n=O
is the default condition

TI:

User terminal

CR:

Card reader

LP:

Line printer

TTO:

Operator console

DB:

RPJ04/RPJ06 disk

DP:

RP11/RP03 disk

DM:

RK06 disk

MM:

Magnetic tape

MT:

Magnetic tape

* See your System User's Guide for
specific device types.

2-4

CHAPTER 3
CONTROLLING SORT OPTIONS

3.1

SWITCHES

SORT switches describe the input and output files and allow you to
change SORT defaults, such as the number of scratch files used. RMS
requirements have made it necessary to completely revise the SORT
switches. These changes have been reflected in the switches described
here. All numeric switch values are decimal, not octal. All switches
are valid for both input and output files and on all systems unless
otherwise noted.
Table 3-1 summarizes the switches and
other
pertinent information for your convenience.

3.1.1

ALLOCATION (/AL:n)

The ALLOCATION switch is valid for the output file only. You can use
this switch to specify the initial space allocation for the output
file. Legal values range from 0 to 65,535. If you do not use this
switch, the output file allocation defaults to the input file size for
the record and tag sort processes. For index and address routing sort
processes, the default size is based on the number of records sorted.
See Section 5.1.4 for details.

3.1.2

BLOCKSIZE (/BL:n)

The BLOCKSIZE switch is valid for magtape files only and is used to
specify the nonstandard magtape blocksize.
If this switch is not
used, the blocksize defaults to the standard 512-byte block.

3.1.3

BUCKETSIZE (/BU:n)

The BUCKETSIZE switch specifies the RMS bucket size (the number of
512-byte blocks per bucket}. Please note that the size of the block
is the standard 512 bytes even if the BLOCKSIZE switch is used.
The
default value of this switch depends on the organization of the input
file. If the input and output files are of the same organization, the
default for the output file is the input file value. If the input
files differ in organization or the default is required for the input
file, the default value is 1.

3-1

CONTROLLING SORT OPTIONS
Table 3-1
Switch Summary
Switch Name

Code

Default Value

Validity

ALLOCATION

/AL:n

See Section 3.1.1

Output

BLOCKSIZE

/BL:n

512 bytes

Both
(magtape)

BUCKETSIZE

/BU:n

Input file or 1

Output

CONTIGUOUS

/CO

Non-contiguous

Output

DEVICE

/DE:x

Build value

Input

FILES

/FI:n

Build value or 5

Both

FORMAT

/FO:x:n

(See footnote)

Both

INDEXED SEQUENTIAL

/IN:x

(See footnote)

Input

KEY

/KE:abm.n

C, N, first position
of field, length of
field

Input

PROCESS

/PR:x

R (record sort)

Input

RELATIVE

/RE

(see footnote)

Output

SEQUENTIAL

/SE

(See footnote)

Output

SIZE

/SI:n

RSTS clustersize or
RSX-llM retrieval
window size (see
Section 3.1.13)

Output

--····--··-·----- -·-·---------··

--···-------···---·---····----··-

--- - - - - - - - - -

----

* The default for input files is the type of
time.

file

present

OPEN

The default for output files is SEQUENTIAL. For SORTR and SORTT,
the
output record format will default to the characteristics of the input
file unless you use STREAM ASCII input record format.
STREAM ASCII
will produce a VARIABLE output record format.
For SORTI and SORTA,
the output record format will be FIXED.
Default record sizes are six bytes for SORTA, six bytes plus the size
of the index for SORTI, and the size of the input record for SORTR and
SORTT.

3-2

CONTROLLING SORT OPTIONS
3.1.4

CONTIGUOUS (/CO)

The CONTIGUOUS switch is valid for output files only.
It specifies
that the output file will be contiguously allocated. This means that
each successive block assigned to a file will be physically located
between its logical predecessor and its logical successor with no
filler or extraneous material separating the blocks. The default is
non-contiguous.

3.1.5

DEVICE (/DE:x)

The DEVICE switch lets you specify which device you want to be
associated with the scratch files. This switch overrides any device
specification from task build options. The parameter x is any valid
1- to 4-character device name with the colon that normally follows
that name omitted (e.g. DB3). See Section 5.1.6 for details.

3.1.6

FILES (/FI:n)

The FILES switch specifies the maximum number of intermediate scratch
files.
Legal values are from 3 to 8. If you patched the location
FILES at task build time, the default value now is the one you
specified then. If you did not use a patch, the default is 5.

3.1.7

FORMAT (/FO:x:n)

The FORMAT switch specifies the record format (x) and the maximum
record size (n) in a file. The record format x may take the following
values:
FIXED,
STREAM,
VARIABLE, or
UNKNOWN
Record format values may be truncated to the first letter.
The
maximum record size n is the exact record size in bytes for FIXED
length records. For the other formats, n is the size in bytes of the
largest record present. The record size is required only on input.
This switch must be present in input files even if the format is
UNKNOWN.
The default of the record format (x) is VARIABLE. The
record format on output can be used to override the input record
format. For the default of n, see the footnote of Table 3-1.

3.1.8

INDEXED SEQUENTIAL (/IN:X)

The INDEXED SEQUENTIAL switch specifies the INDEXED SEQUENTIAL file
organization.
For default values, see the footnote of Table 3-1. X
specifies the number of keys (no default). This switch is valid on
input only.

3-3

CONTROLLING SORT OPTIONS
KEY (/KE:abm.n)

3.1.9

The KEY switch is the most important switch in the SORT parameters.
It tells SORT wh.ich fields are to control the sequence of the output
file.
You can specify more than one key field
(up to 10)
if you
separate each description from the next with a colon. Here is an
example of two key fields:
/KE: BN 1. 6: C 8. 2

The field description sequence in the
which breaks down as follows:
a

basic

format

above

abm.n

Specifies the way in which the data is to be handled. If it is
omitted, the default value is C as described below. If you have
not used SORT before, refer to Appendix B and become acquainted
with the concepts of zones and digits.
Here are the legal values for the data handling parameters and
their interpretations:
B

2's complement binary

Alphanumeric

If the characters are alphabetic, numeric with the
sign superimposed over the units digit, or contain
slashes (/), use the value of the digits group (see
Table B-1). Here are two examples and their values:
A2CD5

= (+)

12345

A/47J

= (-)

11471

If the characters represent a standard FORTRAN IV
number, such as 12, -35, 42.98, or -0.76E+3, convert
the number to binary for storage or evaluation

2- or 4-word floating point binary

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

Same as I but with the sign trailing and separate

Same as D but with the sign leading and overpunched (54321,
for instance, if positive, would come out as 5432A. The
negative of 54321 would be 5432J.)

Packed decimal format

ASCII zone (see Table B-2)

Defines the general sort order.

The default is N (ascending order)

Ascending order

Opposite, or descending, order

Is a decimal number giving the first byte of the key field. Number
from the first byte of the record which is byte 1. This item must
be present.

Is a decimal number giving the length of the key field in bytes.
This item must be present.

3-4

CONTROLLING SORT OPTIONS
3.1.10

PROCESS (/PR:x)

The PROCESS switch specifies the
Legal values are:
R T A I -

3.1.11

sorting

process

use.

switch, for output only, specifies RELATIVE
For defaults, see the footnote of Table 3-1.

file

RECORD (the default)
TAG
ADDRESS ROUTING
INDEX

RELATIVE (/RE)

The RELATIVE
organization.

3.1.12

type

SEQUENTIAL (/SE)

The SEQUENTIAL switch, for output only, specifies SEQUENTIAL
organization. For defaults, see the footnote of Table 3-1.

3.1.13

file

SIZE (/SI:n)

The SIZE switch specifies the file cluster size (RSTS) or the size of
the retrieval window (RSX-llM).
The ranges and defaults for this
switch will be determined by your operating system. Check with your
system manager for this information and record the ranges and default
below:
SYSTEM:

3.2

~~~~~~~

RANGE:

~~~~

DEFAULT:

~~~~-

THE SPECIFICATION FILE

The specification file offers a variety of controls for the sorting
process.
These controls are entered in a format consisting of up to
three types of records or lines in the specification file;
a fourth
type may be used if nonstandard collation is required. Figure 3-1
shows, in card form, two possible arrangements of the specification
file whose elements are explained below.

3-5

CONTROLLING SORT OPTIONS
.--------------·-------------------------··--~-------

DATA FIELD

····~---·

FIELD

---

"""'""o"
- ·-RECORD TYPE
SPECIFICATION

FIELD
ECIFICATIONS

DATA FIELD

RECORD TYPE

-----------·-,I

FIELD
SPECIFICATIONS

ALTSEQ

I
I

HEADER

ALT-;E;- -

HEADER

RECORD TYPE
SPECIFICATION

t--ALTERNATE COLLATING
SEQUENCE RECORDS

L-ALTERNATE COLLATING
SEQUENCE RECORDS

HEADER CARD

With Record Type Specifications

Figure 3-1
1.

FIELD
SPECI Fl CATIONS

Without Record Type Specifications

Specification File Format

The Header
The first record in a specification file must be the
The header tells the SORT program:

Header.

•

the kind of sorting process to be used,

•

the key field and output record size,

•

whether an alternate collating sequence is
and

•

whether the key fields are to be stripped from
output. There is only one Header for each job.

used,
the

ALTSEQ Records
If you use an alternate collating sequence, ALTSEQ records
follow the header.
ALTSEQ records allow you to change the
order in which a character or characters is sorted. You can
use these records to:

3-6

CONTROLLING SORT OPTIONS

Move certain records to the front or back of a file

Group them in one place within the file

Change the order in which records appear in a
sequence within the file

given

Record Type Specifications
If you do not use an alternate collating sequence, the next
type of record or line in the specification file is record
type specifications. They control record selection and allow
SORT to work with variable record formats.
If all records
have an identical format (or the format is not important) and
all are to be included in the output file, the record type
specification may be omitted.

Field Specifications
If record type specifications appear, each is followed by any
applicable
field
specifications.
If
not, the field
specifications follow the header (behind any ALTSEQs).
The
two
kinds
of
field
specifications
are
key
field
specifications and data field specifications.
Key field
specifications define and locate each key. They are listed
in
order
of
decreasing
significance.
Data
field
specifications define and locate all the data fields to be
written to the output file. The data will appear in the same
order in the output file as the data field specifications.
Key field specifications appear first in the specification
file and are followed by any data field specifications. Data
fields are not specified for SORTA or SORTI because the
output
file
contains only pointers and possibly some
restricted-format key data.
The format of each type of specification record is fixed.
The SORT specification form, based on card columns, is shown
in Figure 3-2. The following entries are common to all three
types of specification file lines.
Column

Entry

Notes

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
Specifications. Only the first
page has a header specification.

3-5

Line number

Depicting 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, I, O, or F as shown below

40, etc. Comments

Ignored by SORT processing

3-7

CONTROLLING SORT OPTIONS

rn
1

SORT SPECIFICATIONS
Page

HEADER SPECIFICATION

Program

ldent1f1cation

ffYfTTI

Programmer

SORTR,T

Mode of Processinl)
0

SOR TR
SOR TT
SORTA
SORTI

Line

Total
Length of
Key Fields

RECORD TYPE SPECIFICATION

.......----- -.-----Factor 1

;c u . 1 - - - - - - - I

Lme

g §~
~

:; Ci

~ ~ N

~ ~ u

Field Location

~-~~~---------- ,_

Factor 2

1-.._.------------

~~ f-

------~----

Constant -

1-- --:- -

GE
LT

-_ - -,

Field Location

-------.-- -----------·----·---·--·--·

- -

Comments

FIELD SPECIFICATION

- - - - - - - - - - - - - - - --------------------· --------------------

Forced

J
s
0

~::_ ~

0 ;;:

Field Location

~
: 6 g

I-----,

H1---F-ro-m~-To--~H j
1

FiekJ Name

Comments

I
I

a 9 10 11 •-~- 1!}~ -~s ~+1~~-\~~ 21 ~~ ~J 2\.~~-.~~ ~-! _:_s ;iq rn 11 J? 3-~- -~~~ .~s 1_ri_ -~1 ~~ :? 4~ 41 42 43 44 45 46 47 48 49 so~-\~ s.1 _:_~ ~-~ s.& ~~~-~~ ~.3.- ~!,2..~. ~s 66 67 68 69 10 11 12 13 14

f--1
0 8

f-+

f-·

-·

l- -

1 0
1

f-+-1 2

1 3
1 4

Figure 3-2

SORT Specification Form
NOTE

Whole lines of comments may also be
included in the file; these are denoted
by an asterisk (*) in column 7.
The
comment lines may appear anywhere in the
specification file and serve to document
the file in its printed listings. They
do not affect SORT operations.
In the following material,
criteria apply:

unless

otherwise

stated,

the

following

•

Numeric data is decimal, not octal.

•

Either leading zeroes or leading
right-justified entries.

•

All field position definition records begin at column 1.

3-8

blanks

are

acceptable

CONTROLLING SORT OPTIONS

Header Specification

3.2.1

See Section 3.2.1.1 for notes and comments:
Columns

Explanation and Legal Entries

1 -

Page and line numbers: described in
Section 3.2

Header record identification

Legal value: H
7 - 12

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

13-17

Total of all key field sizes
Legal values: 1 - 16383
N.B.: 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

19 26

(not used)
Alternate collating sequence
Legal values:
blank - standard ASCII sequence
E - EBCDIC sequence
X - user-modified ASCII sequence

(not used)

Output option for SORTR and SORTT ONLY
Legal values:
X -

key fields created by SORT will
be stripped from the output record
blank - no action
29-32

Output record length for SORTR and SORTT
ONLY
Legal values:
decimal number equal
to the
number of bytes in the largest
output record
3-9

CONTROLLING SORT OPTIONS
Columns

Explanation and. Lega~_ En_tries

33-39

{not used by SORT - may be used for
comments)

3.2.1.1

Notes and Comments on Header Specification Entries
Notes and Comments

Column of
Entry Affected
18

This

field may be qualified by N or
O entered in column 7 of the field
specification (q.v.)

This field identifies the sort order for
fields identified as alphanumeric {C
in
column
8
of
the
field
specification).
If X is entered
here, the header must immediately be
followed by ALTSEQ
cards.
See
Section
3.2.1.2 for ALTSEQ card
format. Standard ASCII is assumed
for forced key field specifications
and this field cannot be used to
amend that assumption.
SORT does
not perform file transliteration.
Files
that
are not already in
standard ASCII characters must be
transliterated before the sort.
The EBCDIC
characters
sequence.

option specifies ASCII
sorted
into
EBCDIC

If a blank appears here, the key fields
may be treated as data fields and
moved to the output record as any
other data field would be, whether
relocated within the output record
with respect to the input record or
not.

29 - 32

To determine this number, add the sizes
of all the data fields in the field
specifications {plus the sizes of
the key fields if column 28 is
blank) for the largest record in the
file. If neither SORTT nor SORTR is
to be run at this time, then an
entry in this field is not needed.

3.2.1.2

ALTSEQ Format and Notes

Column

Notes

1 - 6

ALTSEQ - mandatory entry

7 -

{not used)

3-10

CONTROLLING SORT OPTIONS
Columns
9 - 80

3.2.2

Notes

Replacement character groups {12), of 6 positions each;
each group contains the octal representations of
two ASCII characters: the character being sorted
out of
sequence,
and
the
character
that
corresponds to its new position in the sorting
sequence.
For example, the correct entry to
replace an ASCII space with an ASCII zero (0)
would be 040060; 040 for the space and 060 for
the zero.
The result of this would be to sort
spaces into the same relative position in the
sequence as Os.
Notice that preexisting Os are
not affected by this entry, but they may be
affected by another entry on the same or another
ALTSEQ card under the same header.

Record Type Specification

See Section 3.2.2.1 for notes and comments.
Columns

Explanation and Legal Entries

1 -

Page and line numbers. See Section 3.2
for description

Inclusion or Omission character
Legal values:
I - Include the records described
in this line in the sort
O - omit the records described
in this line from the sort

Continuation character {allows for cases
in which several conditions define a
single record type)
Legal values:
A - logical AND
0 - logical INCLUSIVE OR

Field mode {defines the processing to be
applied to the data field defined}
Legal values:
B C D F I J -

KP Z -

binary
alphanumeric
numeric
binary floating point in
2 to 4 words
numeric with leading and
separate sign
numeric with trailing and
separate sign
numeric with sign leading
and overpunched in first byte
packed decimal
zone, see Table B-2, Appendix B
3-11

CONTROLLING SORT OPTIONS
Columns

Explanation and Legal Entries

9 - 12

Factor 1 (position of the base comparison
factor is in the first byte; the first
byte in the record is byte 1)
Legal values:
a decimal number - location of the
first byte
blanks - specifies a single byte
field

13 - 16

Factor 1 endpoint
Legal values:
a decimal number - location of
last or only byte in the
field

17 - 18

Relationship of Factor 1 to Factor 2
Legal values:
EQ NE LT GT LE GE -

equal
not equal
less than (i.e. Fl LT F2)
greater than
less than or equal to
greater than or equal to

Comparison factor format (indicates whether
Factor 2 is a field or a constant)
Legal values:
F - identifies a field
C - identifies a constant

20 - 39

Factor 2 (This takes two different formats
depending on whether Factor 2 is a field
or a constant.)
If Factor 2 is a field:
20 - 23

location of first byte - a
decimal number or blanks
as for Factor 1.

24 -

location of the last or only
byte - a decimal number as
for Factor 1.

28 - 39

(not used)

If Factor 2 is a constant:
20 - 39
40 - 80

the value of the constant

(not used - available for comments)
3-12

CONTROLLING SORT OPTIONS
3.2.2.1

Notes and Comments on Record Type Specifications

Columns

Explanation

{general)

If all the records in the input file are to be
included in the sorted output file and they all
have the same format or you are using the entire
record
as
the
key,
then
no record type
specification is necessary. If all the records
have the same format and some are to be omitted,
then only one record type specification is needed.
If several different formats of records are to be
sorted, then there must be
a
record
type
specification
for
each.
Each
record type
specification used should be followed by its
appropriate field specifications.

If this column contains an I and there are no
further entries in this line, all records not
previously described are to be included in the
sort input.
Records that are not described in
Include lines will be ignored.
Because SORT
processes Include and Omit lines sequentially, you
should be very careful to get your lines in the
right order, particularly if any records are
described on more than one line. The last Record
Specification line before a Field Specification
must be an Include line.
If you describe the
records to be processed by omitting the unwanted
records from the file, then you can use a line
with an I in column 6 and no further entries in
the line as the final record specification.

Multiple AND and OR lines are permissible, but the
series will be processed strictly in sequence,
each line ANDed or ORed to the accumulated Boolean
expression in turn:
If A is False and B and C are True,
the expression A OR {B AND C) is
True logically, but in order to
get
the correct results from
SORT, arrange it this way:
B AND C OR A
SORT will react:
{B AND C) OR A,
which gives
value.

the

required

True

On the other hand,
A OR B AND C
gives a True value in both the
SORT and non-computerized logic
as it stands.

3-13

CONTROLLING SORT OPTIONS
Columns

Explanation

C - alphanumeric information limited to 256 bytes
D - numeric value, maxi mum size 8 bytes after
conversion to binary.
If the number is
presented in standard ASCII Fortran
IV
format, as for a decimal number with or
point,
without decimal
floating
point
exponent, or sign at the beginning, the SORT
will convert the number to binary. But the
maximum size is still 8 bytes.
F - maximum size 8 bytes - two or four
word
floating
binary
representation
before
conversion

z - ASCII zone, maximum size 1 byte

see

Table

B-2 in Appendix B
9 - 16

Factor 1 - Sort identifies records by comparing
the value of one or more fields in a record
with constants or other fields in
the
record. These other fields or constants are
referred to as Factor 2.

17 - 18

defines the Factor 1 - Factor 2 relationship

20 - 39

Factor 2 - See notes under Factor 1 above.
If
Factor 2 is a field, it must be the same
length as Factor 1. If a constant, after
conversion to the corresponding internal
format, be the same length as Factor 1 and
be
in the same mode (see Field Mode,
position 8).

3.2.3

Field Specifications

There are normally a set of field specifications following each record
type specification in the specification file set.
(For one exception,
see paragraph 3.2.2.1, general notes.) If there are no record type
specifications,
any appropriate field specifications follow the
header. Field specifications have two main jobs:
•

to describe the output file format

•

to determine, for each record type, the
sort is to be based.

keys

which

the

List the key field specifications first,
in order of decreasing
significance.
Then, if the sorting process is SORTR or SORTT, add
data field specifications to put out any portions of the original
input record.
List the Key and Data field specification lines in the specification
file in the same order as the required output record format. The SORT
creates the output record according to the order, size, and contents
of the key and data specification lines. Place an X in column 28 of
the Header to strip unwanted key fields from the output record.
If
the key field is specified twice, as a key field and as a data field,
the key field ·can remain in the output record where the data field
specification puts it, even if the key fields are stripped. See
Section 3.2.3.l for further commentary.
3-14

CONTROLLING SORT OPTIONS
Columns

Explanations and Legal Entries

1 -

Page and line numbers - see Section 3.2

F - specifies a field specification

Field type - specifies keys and their sort
sequences and data fields
Legal values:
N - key field, normal sort
sequence (see position 18 of
Header)
O - key field opposite sequence
F - key field, special treatment
required (see positions 17 - 19)
D - data field (if used, then
column 8 should be C)

Field type - should be C if column 7 is
D - these options are the same as those
for column 8 of the record type specifications (q.v.).

9 - 12

Field location - location of first byte of
multi-byte field
Legal values:
a decimal number - location of first
byte of field
blanks - specifies one-byte field

13 - 16

Field location - last or only byte
Legal values:
a decimal number - location of last
or only byte in the field

17 - 19

Forced key fields
17

Trigger character - identifies the
character needed to trigger the
force:
Legal values:
blank - the force always occurs
any other character - the force
occurs only if this character is
found

Replacement character
Legal values:
any legal character
3-15

CONTROLLING SORT OPTIONS

Columns

Explanati~n

and Legal Entries

Continuation character
Legal values:
blank - this is a new forced
key
any other character - this
key is continued from
the previous line

20 - 80

3.2.3.1

{not used - available for comments}

Notes and Comments on Field Specification Entries

Columns

Explanation

Enter key fields first in order of
decreasing significance.

17 - 19

Use this field to change the content of
a I-character key field, thereby
changing the sorting sequence. The
original data will still appear on the
output record, since only the key will
be changed. To specify a forced key,
put F in column 7 of the field
specification and the nature of the
force in columns 17 - 19. Since the
force is based on characters, there must
be a C in column 8. A force does not
apply to any other keys or fields than
the one stated.

To help you understand how these forced key fields affect the
here is a description of how the SORT handles these fields:

SORT,

When column 7 contains an F, SORT sets up a one-byte field in its key
area and sets its value to either 377{octal} for ascending or 0 for
descending sequence. Note that the alternative collating sequence
does not apply to forced key characters.
If column 17 is not blank, SORT compares its contents to the content
of the field defined in columns 13 - l~.
If an equal condition
occurs, SORT moves the character to the reserved key byte previously
described.
By using more than one Field Specification, you can force more than
one character out of normal sequence, such as sorting the entire
alphabet out ahead of numerics. But beware:
1.

Unless the displaced character is itself displaced to another
position in the sorting sequence, it will appear with the
character displacing it in the order in which the two
characters were found in the input file.
Thus, if A
displaces ! and a series appears in the input file as {A, B,
!,
A, !, !, A, C, D}, the sorted output wi l 1 be {A, ! , A, ! ,
! , A, B, C , D} , so ...

When in doubt, force the displaced character elsewhere,
particularly when it is known to appear in the input file and
is not omitted elsewhere in the SORT specification file.
3-16

CONTROLLING SORT OPTIONS
If, on the other hand, column 17 contains a blank, no comparison is
performed and the value of column 18 always replaces the initial value
in the reserved key byte. This has the following results:
1.

When associated with conditional keys in preceding lines,
records not satisfying the prescribed conditions can be
grouped into any position in the output file, not necessarily
the end of the file.

When this specification stands alone,
it
becomes
an
unconditional force of all records in the group. This form
can be useful when the group is already identified by a
record type specification.
It has no effect on the entire
file if applied to the entire file.

If column 19 of the field specification is blank or the preceding line
does not specify a force key, SORT only reserves a new key byte. The
same byte is modified in all other cases. This makes it possible to
combine several conditions within the same key, such as arranging an
output file in the order 4, 2, 3, 1, rather than 1, 2, 3, 4 or 4, 3,
2, 1. However, SORT still expects the field location to be defined in
columns 13 - 16, even if each continuation line covers the same field.
NOTE
SORT rejects forced
for SORT!.

3.2.4

key

specifications

Using a Specification File

The sample input file contains records in the format shown in Figure
1-1.
The file may contain null records such that there may be a
record identifier and no other data in the record.
SORTR is entered in columns 7 through 11 of the Header.
The header
specification also indicates that the key is to be sorted in ascending
sequence and that an alternate collating sequence is to be used.
The
ALTSEQ line following the header tells SORT to interpret all nulls as
spaces.
The null records are omitted from the sorting process with an Omit
line in the record type specification. The Include lines direct SORT
to process only the RESTOCK records where the present amount of stock
is less than the stock quantity allowed.
ORDER Records, SALES
Records, and RESTOCK records where the present amount of stock is
greater than the stock quantity allowed are ignored.
The field specification shows that the selected records will be sorted
on the reorder number. Since this is a SORTR, the Item Number Code,
Unit Code, and Cost per Unit data fields, in positions 2-14, transfer
to the output records.
The output records will have the reorder
number in positions 1-6, and the Item Number Code, Unit Code, and Cost
per Unit in positions 7-19. Note that the total length of the key
fields and the output record length in the header specification are
calculated from the field sizes indicated in the field specification.

3-17

CONTROLLING SORT OPTIONS

110•00•0
~ ~
~ ~~~~::~~.~~Is~~
E Q U I PM E NT

SORT SPECIFICATIONS

75 76 77 78 79 80

i~~;~1:a1ion lslTIKlo§k I

Page~

Date _ _ __

HEADER SPECIFICATION
Programmer ------~

RECORD TYPE SPECIFICATION
~ "-

Factor 1

g g~
~ ~~
5

t:. --:" ___ ~onstant

From

.,M

Locatio~
Frori
To
I
F 1elcJ

1
75;•&!<1, sj;

.·~

f-~

Record Name

Comments

,~ v[e~~

•it•Yµ:i°CY •trl

f '''''

-Hf--_,-

-+-·

1+1

1---1

-1-j

0 5
0

-----·---t_ ____

I' • ."••!'>I.

- - - · --·

11 1s.__~20 212_! -~2+-~26r~~ 2:!~!2_2 33 34 JS 36 37 3~ 39 40 41 4243 44 45~~ 47'~~~h51 52 53 54 55 56 57 58 59 60 6162 63 64 &s 66 67 ~a 69 1~ 11 n 13 74

9 IO 11 12 13 14 15 l6

111 1J>

f--+~

Factor 2

Field Location

H
0

~-~~~·-----·-·-·-···

I
fl
I
FIELD SPECIFICATION
- ---·,.------,' ----------·------------· -----, ---·------ ..--------------------------Forced

.._

Field Location

From

~ ~

~ 2 0 ·--~----1~0~- n_
~ -

~ lt ~-

? u

9 10 II 12 13 14 15 16 17 I! 19 2~ 21 ~~ 2! 14 25

Fri~1

F~c

o 8
0

f-1

NOT USED

~-----,
Field Name

Comments

I
I

26 27 28 ~? ~~~ ?~2· Jl J4 ~-~ 3~ J7 38 39 40 41 42 43 44 45 46 47 ~- 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74

~C.t!l4S~

I~~ ~

Pi~Xo~~i(,~~~IMN ~u~~vDI

~K~

1 0

1---1

1 2

1---

-~+H

t-·

-t--+- - f-j--1- t -

1 3

t--

1 4

Figure 3-3

SORT Specification Entries

3-18

1---+-

tt-

- f-+-+- -1--+- +-- ----

t-+--

CONTROLLING SORT OPTIONS
3.3

SORT SPECIFICATIONS SUMMARY

Table 3-2
Header Specifications
Column

Entry

Explanation

1 - - - - - - - !------------------------- t-----

---------~

Header Specification

7-12

SORTR
SORTT
SORTA
SORT I

Record Sort
Tag Sort
ADDROUT Sort
Index Sort

13-17

1 - 16383

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

A or blank

Processing sequence: Ascending or
Descending

Blank
E

29-32

ASCII collating sequence
EBCDIC collating sequence
ASCII with modification listed in
following lines

Blank
(SORTR, T
only)

The key fields will reside at the
beginning of each output record

The key fields will be stripped from
the beginning of the output records

Decimal Number
(SORTR, T
only)

This entry specifies the number of
bytes of all key and data fields
listed in the Field Specifications
(must be maximum for SORTR)

3-19

CONTROLLING SORT OPTIONS

Table 3-3
Record Type Specifications
Entry

Column
6

Explanation

The records described in this record
specification will be included in the
sorting process. If there are no entries in the following columns, all records not yet described will be included

The records described in this record
specification will be omitted from the
sorting process

The condition identified by this record specification is a logical AND with
the condition from the previous line

The condition identified by this record
specification is a logical OR with the
condition from the previous line

Character - the field in columns 9-16
(Factor 1) is alphameric

Zone - use the zone value for the entry

Digit - use the digit value or convert
the FORTRAN IV number to binary

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

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

Same as D but with sign leading overpunch, i.e., sign is superimposed upon
first byte of the field

Packed - the field contains packed decimal data

Binary - the field contains 2's complement binary data

Floating - treat the field as a 2- or
4-word floating binary representation
(Continued on next page)

3-20

CONTROLLING SORT OPTIONS
Table 3-3 (Cont.)
Record Type Specifications
Explanation

Entry

Column

- - - - - - - - - --- - - - - - - - - - - - i

Decimal
Number

Locati on of the first byte of the field
(Facto r 1)

Blanks

Single byte field

13-16

Decimal
Number

Locati on of last byte of field or location o f single byte field

17-18

EQ
NE
LT
GT
LE

Factor 1 must equal Factor 2
Factor 1 must not equal Factor 2
Factor 1 must be less than Factor 2
Factor 1 must be greater than Factor 2
Factor 1 must be less than or equal to
Factor 2
Factor 1 must be greater than or equal
to Fae tor 2
The f allowing entry describes a field to
be co mpared with the field specified in
col um ns 9 through 16

9-12

GE
19

The f allowing entry describes a constant
to be compared with the field specified
in co lumns 9 through 16

Decimal
Number

Locat'ion of the first byte of the field
(Fact or 2) compared with Factor 1

Blanks

Facto r 2 is a single byte field

24-27

Decimal
Number

Posit'ion of last byte of Factor 2 field
or lo cation of single-byte field

20-39

Character
String

The c onstant compared with Factor 1 entry must agree with mode and length of
Facto r 1

20-23

3-21

CONTROLLING SORT OPTIONS

·rable 3-4
Field Specifications
··- .....- -..- ..

Entry

Column

- -----·-··-···--- --- ··-------·------ -----~

Explanation

Field Specification

Normal - key field sequenced as indicated
in column 18 of header

Opposite - key field sequenced opposite
to column 18 of header

Forced - single byte key field with
special sequence

Data field

Character - alphanumeric key or data
field

Zone -

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

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

Same as D but with sign trailing and
separate, i.e., the last byte of the
field is + or a -

Same as D but with sign leading overpunch, i.e., sign is superimposed upon
first byte of the field

Packed - the key field is in packed
decimal

Binary - the key field is in 2's complement binary notation

Floating - the key field is in 2- or 4word floating binary representation

9-12

Decimal
Number

Location of first byte of field

Replacement
Character

The character which will be substituted
when the force condition {if any) is met

Continuation
Character {any
character other
than blank)

This specification line covers the same
forced key field as the preceding line{s)

Blank

This line designates a new forced key

zone value

________.

---·---·-----·-·---- .............

3-22

CHAPTER 4
ERROR CONDITIONS

4.1

ERRORS IN SORT: GENERAL COMMENTS

There are three main categories of errors that SORT can detect:
•

Command decoder errors

•

Specification file errors

•

I/O errors

SORT errors are reported at the terminal only and are not recoverable.
The format of a SORT error is as follows:
SRT -- control-phase:?message

[-RMS-status-code]

where:
control-phase

is the SORT phase in control at the time
error occurred. Legal values are:

the

c - command decoder
M - merge
p - presort

4.2

message

is a one-line
happened.

RMS-status-code

is a decimal status code returned by RMS for
additional information on file errors only.
If RMS is not impacted by the SORT error,
this status code does not appear. Status
codes likely to be seen are listed with their
meanings in Section 4.7.

brief

explanation

what

COMMAND DECODER ERRORS

If your error message has a control-phase of C, you have made a
command decoder error. The command decoder error messages and their
probable causes are listed below:
1.

SORT COMMAND ERROR
a.

Too many input files
(more
than
two,
including
specification file) or output files (more than one)

General syntax error

4-1

ERROR CONDITIONS

Too many switches

Erroneous switches on the specification file

An undefined switch

IMPROPER SWITCH:

/FI

Less than three or greater than eight scratch files.

Invalid terminator
NOTE
Valid terminators are period, comma, slash, equal
sign and <CR>. "INVALID TERMINATOR" means that
some other character was used as a terminator or
that SORT expected to find a terminator where
none existed.

IMPROPER SWITCH:

/KE

Invalid letter or value

Start location or size is 0

No period (.) between start location and size

Illegal size for data mode

Invalid terminator (See NOTE above)

TOO MANY KEYS
Buffer space overflowed
NOTE
SORT reserves a buffer area for storage of a table
based on the input specifications in order to control
the processing of each record. This space should be
ample for all situations to make this error unlikely.
If you need more space, see Appendix D.

NO KEYS SPECIFIED
There are no key switches in the command
specification file has been declared.

string

and

KEY AFTER LAST BYTE OF RECORD
The end of an input record key field goes
record size (switch or specification).
4-2

past

the

stated

ERROR CONDITIONS
7.

NO /FO SWITCH
You omitted the /FORMAT switch on the input file.

IMPROPER SWITCH:

/FO

You have not specified a valid format type.

IMPROPER SWITCH:

/PR

You specified an invalid sort process.

4.3

SPECIFICATION FILE ERRORS

In general, the detection of an error
in the
results in an appropriate message followed by:

Specification

•

a printout of the entire record, or

•

a printout of the record up to the point where the
detected.

INVALID CHARACTER

error

Column 6 is not H,I,O,F and record is not ALTSEQ.

Process is not SORTR, SORTT, SORTA, SORTI, or blank.

Collating sequence is not blank, E, or X.

Data type is not B, C, D, F, I, J, K, P, Z.

Key type is not D, F, N, 0.

Logical entry is not A, O, blank, or

File

was

ILLEGAL FIELD
a.

A numeric field in specification
decimal digits or blanks.

No key size is given in Header specification.

No output size is given in Header Specification
of SORT is SORTR or SORTT.

ALTSEQ is misspelled.

ALTSEQ entries do not represent 7-bit octal values.

Last location is less than first location in record field
identification.

Size is invalid for data mode.

4-3

contains

other

than

type

ERROR CONDITIONS

Sizes of Factors 1 and 2 in Record Specification
match.

Compare relation is undefined.

Forced field is other
position.

than

one

Constant given in Factor 2 is greater than 20 characters.

Mode of constant does not agree with mode of Factor.

Invalid characters appear in constant
if the constant is numeric).

Sign is omitted from binary or packed constant.

(e.g.,

non-digits

NO HEADER
First record of
specification.

specification

file

not

INCORRECT SEQUENCE
a.

Numeric record sequence is lower than sequence previously
encountered.

No valid data specification appears when keys are
stripped from output.

Record specification after
should be last).

Key specifications appear after data specifications.

"include-all"

("include-all"

NO ALTSEQ
a.

type

not

ILLEGAL CONSTANT

than

Specification for alternate collation entered in
column 26 but no ALTSEQ Specifications follow.

Header

TOO MANY SPECIFICATIONS
a.

Number of specifications for a particular type of
have overflowed the buffer space.

4-4

record

ERROR CONDITIONS
NOTE
SORT reserves a buffer space for storage
of
a
table
based
on
the input
specifications and used to control the
processing of each record type. This
space should be ample for all situations
to make the error unlikely other than in
very exceptional circumstances.

LAST RECORD SPEC NOT "I"
a.

Last record specification in file was OMIT.

ILLEGAL KEY
a.

A forced field specification was
SORT (SORT!).

included

Index

10. NO KEYS SPECIFIED
a.

4.4

No key specification appeared before data
in a specification file set.

specifications

I/O ERRORS

Once the sort is under way the only normal errors that can
due to I/O failure, as indicated below:
1.

are

xxxx

OPEN (IN/OUT) FAILURE ON filename.ext xxxx
The file could not be opened by RMS.
the appropriate RMS error code.

occur

The parameter

INPUT/OUTPUT ERROR ON filename.ext xxxx
RMS generated an error while reading or writing this
The parameter xxxx is the appropriate RMS error code.

BAD RECORD SIZE ON filename.ext
RMS tried to read a record on this file that is
the size specified in the /FO switch.

file.

larger

than

NO ROOM IN filename.ext
a.

The storage capacity of the device
while writing the output file.

has

been

exhausted

There are no more clusters
(RSTS/E only).

size

you

specified

4-5

the

ERROR CONDITIONS

INAPPROPRIATE FILE ORGANIZATION
a.

/IN switch was not specified.

You tried to access an indexed file and
was not linked in.

support

TEMP FILE ERROR xxxx
a.

4.5

RMS-llK

An I/O error occurred on a scratch file.
error code.

xxxx is the RMS

CONTROL PROGRAM ERRORS

Pre-Sort Errors

4.5.1

NO DATA IN INPUT FILE filename.ext
a.

There are no records in the input file.

The file is not an RMS file.

INVALID KEY FIELD DATA filename.ext
a.

An inappropriate byte has been found in a field described
as D, I, J, K or Z mode.

Merge Errors

4.5.2

# OF OUTPUT RECORDS DOES NOT MATCH # INPUT

The number of records released to the sort does not equal
the number returned.
NOTE
This is a warning message only.
The
output file contains as many records as
specified in the end-of-sort message.

4.6

OTHER ERRORS

Occasionally, errors may occur due to the rejection of data passed to
the SORTS subroutine package (see Chapter 5). They are reported as:
SORT ERROR -- CODE nn
where nn is the code number returned by a SORTS subroutine
in paragraph 5.3).

4-6

(described

ERROR CONDITIONS
4.7

RMS ERROR CODES

A complete listing of RMS status codes can be found in Appendix G.
The status codes that appear in this section represent errors that you
will be most likely to encounter.
The status codes in Table 4-1 occur only under RSX-llM, and are due to
file primitive failure. Check to see if your file is readable and in
the correct format. If these status codes recur, please send an SPR
to DIGITAL.
Table 4-1
RMS Status Codes for RSX-llM
Code

Brief Explanation

-160

Read error on file header attributes

-176

Write error on file header attributes

-400

Deaccess error during $CLOSE

-520

Device positioning error

-560

Files-11 ACP Enter function failure

-624

File extension failure

-1088

File could not be marked for deletion

-1456

Files-11 ACP Remove

-1520

Error while reading prologue

-1776

File write error

-1792

Error while writing prologue

funct~on

failure

The status codes listed in Table 4-2 may occur under any operating
system.
They appear because the file attributes you gave did not
match those of the existing file.
Table 4-2
RMS File Attribute Mismatch Status Codes
r-------...--------------------------------------.

Code

Brief Explanation

-192

Bucketsize exceeds maximum

-480

Dynamic memory exhausted

-976

Key size equals 0 or is too large (indexed
file) or is not 4 (relative file)

-1168
Not enough room to open an indexed file
____
___________________________
_

,__

__._

4-7

ERROR CONDITIONS
Status code -864 occurs during subroutine usage only.
It indicates
that you did not initialize RMS before control passed to SORT.
Table 4-3 gives other status codes that commonly occur.
Table 4-3
Other Commonly Occurring Status Codes
Code

Brief Explanation

-112

Variable length records on ANSI-labelled
magtape are not ANSI D format

-336

CLOSE function failed (RSTS/E only)

-432

RMS tried to access a deleted record

-448

Syntax error, no such device, or device
inappropriate for operation

-464

Syntax error in directory name

-496

Directory not found

-512

Device not ready

-672

RMS attempted to create an already
existing file

-704

File locked by another user - access
denied

-736

File not found

-752

Syntax error in file name

-768

Invalid file options (also occurs if
you try to extend a contiguous file)

-784

Device full - cannot create or extend
file

-880

Illegal operation - see Appendix G
for examples

-896

Illegal record in sequential file or
invalid count field

-1008

Magtape is not ANSI-labelled

-1024

Logical channel busy

-1040

Invalid logical channel number

-1120

Maximum record size equals O
during $CREATE

-1152

Not at end of file

-1200

Open function failed (RSTS/E only)
(Continued on next page)
4-8

ERROR CONDITIONS

Table 4-3 (Cont.}
Other Commonly Occurring Status Codes
Code

Brief Explanation

-1408

Invalid record address

-1424

Invalid or illegal record format

-1440

Target bucket locked by another task or
stream

-1536

Invalid record in indexed file - file
may be damaged

-1568

Record size error during $PUT or $UPDATE

-1744

Syntax error in file version number
(not a RSTS/E occurrence)

-1784

Device is write-locked

See Appendix F for a more complete explanation of each status code.

4-9

CHAPTER 5

INTERNAL OPERATION

5.1

FILES

User and Scratch File Name Conventions

5.1.1

Input, output, and specification files may have any name acceptable to
the operating system.
SORT generates its scratch files by using the temporary file creation
mechanisms of the operating system under which it is running.
Scratch
files are automatically deleted from your
file directory by the
operating system when SORT is no longer running.

LUN Assignments

5.1.2

The following are the default Logical Unit Number assignments as
reflected in the task build command file SRTxxx.CMD (xxx is system
dependent - see Section F.2):
-

-"""·~-~-~-

LUN

Device

Use

Internal Name
---

---

TI:

Command Input File

TI:

Message Input File

SORT: MSGLUN

SY:

Input File

SORT:INLUN

SY:

Output File

SORT:OUTLUN

SY:

Specification File

SORT:XLUN

SY:

First Scratch File

SORTS: $RFIRL

7-13

SY:

Other Scratch File

To change the internal correspondence between LUN and
the following statement in the command file:
GBLPAT=x:y:z
where:
x:y
z

is the internal name, such as SORT:MSGLUN.
is the new LUN in octal.
5-1

usage,

include

INTERNAL OPERATION

To change the devices to LON correspondence,
statement in the command file:

change

the

following

ASG=x:y
where:
x
y

is the device name.
is the LUN in decimal

File Maintenance

5.1.3

At the end of a successful run, SORT stores the new output file on the
device you specified.
The new output file may then be used as
immediate input to some other program. The input and specification
files are unchanged.

File Allocation

5.1.4

To optimize your sort for speed, SORT tries to allocate in advance the
output file and each scratch file at maximum size. SORT does this by
making an educated guess about the space needed. Under the tag and
record sorting processes, SORT allocates as much space to the output
file as to the input file.
Under the index and address routing
processes, SORT calculates a size for the output file by multiplying
the size of the output record by the number of records sorted.
Each
scratch file is arbitrarily given 50 percent more space than the input
file, no matter what process is used.
Usually, this pre-allocation process causes no problems.
But you
should
be
aware
of the instances in which problems due to
over-allocation do arise and what you can do about them.
Your pre-allocation size may be too large if you are using the tag or
record sorting process and one or more of the following conditions are
true:
1.

You are sorting from an indexed file.

You are using OMIT lines in the specification file.

You are using the reformatting option
file.

You are using a smaller record size in the output
in the input file.

the

specification
file

than

Some wasting of space in the output file can usually be tolerated, but
on a crowded disk you may be notified of insufficient space when there
really was enough to hold your file. Try cutting down on the amount
of space by making a conservative estimate of what you need and
including that estimate in the /AL switch on the output file side of
the command string.
You may perhaps be more concerned about the situation in which you get
the following error message when you try to sort a large file:
P:?TEMP FILE ERROR -784

INTERNAL OPERATION
This means that there is no more room for the allocation of scratch
files.
Here, too, there may be space wasted due to over-allocation.
Read Section D, especially Section D.3, again carefully, and remember
that there is a tradeoff in this situation between size and speed.
Try cutting down the size of the scratch file to one third that of the
input file. If the error persists, trim the size again. If this does
not work, then the disk may be too full to handle the sort.

Scratch File Structure

5.1.5

The scratch files consist of one or more sequenced strings of records,
followed by an end-of-file flag word. Each string consists of zero or
more records, in order by key, followed by an end-of-string flag word.
The number of strings on each file varies from time to time as the
sort progresses, until, at the end of the merge phase, there is one
string on each file. If there is not enough data in the input file to
produce at least one string on each scratch file, the unused files
will contain only an end-of-file flag word.
The format of the records stored in these files is as follows:
Word 1

Size of the data portion in bytes.
The
maximum acceptable to SORT is 16K (up to
37777 octal). Thus, bits 14 and 15 are
always O.

Words 2 through n

After possible conversion to a form more
suitable for logical comparison, the key
fields are stored in reverse.

Words n+l* through m

The data portion for record sort (SORTR)
is the record as read from the input
file; for SORTT and SORTA, it is a
2-word relative record pointer.
This
pointer is joined by the key fields in
their original form during a SORT!.

Two unique records also appear in the scratch files as markers.
is one word only containing all 0 bits except as below:
e

Bit 15

indicates end-of-string flag

Bit 14

indicates end-of-file flag

Each

Scratch Files Use

5.1.6

The scratch files switch, FILES (see Section 3.1.5), is only effective
when operating under special conditions.
For each scratch file
potentially to be used, a certain amount of core is required to
establish internal control blocks. This necessarily makes the sort
take longer by:
1.

Restricting the space available for record storage during the
sort, and

Forcing the usage of scratch files even for small input files

* For SORTR and SORTT controlled by specification file, word n+l is
used to hold an internal control pointer, and the data starts at n+2.
5-3

INTERNAL OPERATION
By specifying a smaller number of scratch files than the default, you
may be able to release sufficient space to do the whole sort in core
and thereby speed up the sort of a small file. The default number is
allocated if the FILES switch is unspecified. This default is defined
at task building time via "GBLPAT=SORT:FILES:X" (3< X <10 octal}. The
SORTS routines use an unbroken ascending series of LUNs starting at
the value in internal location $RFIRL and continuing for the number of
files specified in the internal location FILES (the default is 5).
Depending upon the sorting process and/or the sorting specifications,
each scratch file could require half again as much disk storage as the
original input file. Take this into consideration when assigning the
scratch file LUNs to disk devices. Also, the executive's files system
functions most efficiently if few blocks are occupied and there is
minimal space fragmentation of the disk space. Use private disks if
you can. This will speed up the sort, especially if you put all your
scratch files on private disk space and the input and output files on
public disks. If you have more than one disk, scatter the LUNs among
them so that no two successive LUNs are on the same disk. Use the
fastest disks for scratch files.

5.2

ORGANIZATION OF SORT

As noted in Section 1.2, the SORT program goes through three phases.
Each phase is actually set up as an overlay with a resident control
program to handle the overlays in the correct sequence.
Here are the names and functions of the SORT modules:

5.3

RSORT initializes the system, then serves
error processor for the remaining modules.

SORTC decodes the Command String and Specification File.
This routine is the first section to be called by RSORT.
After SORTC finishes, it is overlaid by SORTP.

SORTP handles the pre-sort phase based upon data transmitted
by SORTC and is associated with appropriate subroutines from
SORTS. It is replaced by SORTM when the pre-sort is done.

SORTM controls the main sort and output.
It
for a new sort when it has finished its task.

SORTS is a collection of subroutines
sorting.

that

the

resident

recalls

actually

SORTC
do

the

SORTS ERRORS

Whenever SORTS detects an error, it returns an octal non-zero code in
the location specified by <Location of Error Code> in the most recent
call. In addition, for I/O errors, Rl contains the address of file
control block for the file in error. The status supplied by RMS is
set into the error status word of that file control block.
(See
Appendix G for error code values.}

5-4

INTERNAL OPERATION
The error codes (octal) and their meanings are listed below:
Error
Code

5.4

Meaning

No errors

Device input error

Device output error

OPEN(IN) failure

OPEN(OUT) failure

Size of current record is greater than maximum size

Not enough work area

RETRN was called after it had exited with a negative
error code (end of sort)

SORT routine called out of order (The order of the
calls should be RSORT, RELES, MERGE, RETRN, ENDS).

Sort already in progress {To do a second sort,
must be called to clean up the first sort).

Key size is not positive, SORTS detected a
negative key size in its calling parameter

Record size not positive

Key address is not even (the keys must start
even address because SORT uses word moves).

Record address is not even

Scratch records will be too large (the size of the
keys plus the size of the largest record must be
less than 37776 octal).

Too few scratch files are given
scratch files must be specified).

minimum

Too many scratch files are given (a
scratch files may be specified).

maximum

End-of-string record was
expected

where

none

was

Unexpected end-of-file

SORT found a record larger than expected

Record length is
SORTI.

not

detected

standard

for

SORTT,

ENDS

zero

SORTA,

USING THE SORTS SUBROUTINE PACKAGE

You can associate the SORTS subroutine package with your own control
program to perform a particular type of sort, provided that the
control program meets the necessary interface requirements.
5-5

INTERNAL OPERATION
Conventions and Standards

5.4.1

All of the entry points in the SORTS subroutines are entered
coding:
MOV
#TAG,R5
JSR
PC,X

TAG: .BYTE
.WORD

N,O
ARGUMENT-!

.WORD

ARGUMENT-N

via

the

where:
PC
X
N

is Register 7,
is an entry label,
is the number of arguments.

All of the sort subroutines require arguments which are word-bound
(16-bit word addresses}.
This is because the routines use word
oriented instructions, rather than byte oriented instructions, for
speed advantages.
A detailed description of the calling procedure for each routine
follows.
In all calling procedures, !ERROR is the starting address of
a user-generated error processor to be used if the call aborts.

5.4.1.1

RSORT - Initializing the Sort
MOV
JSR

TAG: .BYTE
.WORD
.WORD
.WORD
.WORD
.WORD
.WORD
.WORD

-----0
.WORD
.WORD
.WORD
.WORD

#TAG,R5
PC,RSORT

7(11.) ,0;7 (OPTIONALLY 11) ARGUMENTS TO RSORT
<ADDRESS OF ERROR CODE>
<ADDRESS OF NUMBER OF BYTES IN ALL KEYS>
<ADDRESS OF NUMBER OF BYTES IN LARGEST RECORD>
<LOCATION OF MOST MAJOR KEY WORD>
<FIRST LOCATION IN WORK AREA>
<ADDRESS OF SIZE OF WORK AREA IN BYTES>
<ADDRESS OF NUMBER OF SCRATCH FILES>
P T I 0 N A L L Y ----<ADDRESS OF NUMBER OF BUFFERS PER FILE>
<ADDRESS OF CLUSTERSIZE FOR RSTS/RETRIEVAL POINTER
FOR RSX>
<ADDRESS OF PRIMARY ALLOCATION FOR SCRATCH FILES>
<ADDRESS OF FIRST SCRATCH FILE LUN>

For FORTRAN, the call is as follows:
CALL RSORT (IERROR,KEYSIZ,MAXREC,KEYADR,IWRKLO,IWKSIZ,IFILES
[,BIGBUG,SCRCLF,PRIALQ,FIRLUN]}

5-6

INTERNAL OPERATION
where:
KEYSIZ

is the decimal byte count of the total key
size.
KEYSIZ must be even (incremented by 1
if not) and positive (error 12 if not).

MAX REC

is the decimal size in bytes of the largest
record in the input file. MAXREC must be
even (incremented by 1 if not) and positive
(error 13 i f not).
The sum of KEYSIZ and
MAXREC cannot exceed 16,383 (error 16 if
greater).

KEYADR

is the address of the most significant word
in
the
major
key.
KEYADR must be a
word-bound address (error 14 if not).
See
Section 5.4.2 for details.

IWRKLO

is the address of WORKLO, the first word in
the
SORT
work
area.
IWRKLO must be
word-bound (incremented by 1 if not).
See
Appendix D, especially Figure D-la.

IWKSIZ

is the size
(decimal).

I FILES

is the number of scratch files
to
be
allocated. !FILES must be greater than 2 and
less than 11 (errors 17 and 20 respectively
if not) .

BIGBUF

is the number of 512-byte buffers to be
allocated for each scratch file and the
number of contiguous blocks read or written
during each scratch file read or write. The
default is 1. For details see Appendix D.

SCRCLV

is the RSTS/E clustersize or the RSX-llM
number of retrieval pointers. The default is
0. For details see Appendix D.

PRIALQ

is the decimal number of blocks to
be
allocated to each scratch file at open time.
For details see Appendix D.

FIRLUN

is the first Logical Unit Number to be used
by the scratch files. The SORTS subroutines
use the set of LUNs running consecutively
from FIRLUN to (FIRLUN + !FILES - 1) in octal
representation. The default is 5.

the

work

area

bytes

FORTRAN users should remember to designate BIGBUF, SCRCLV, PRIALQ, and
FIRLUN as INTEGERS.
5.4.1.2

RELES - Passing Input Records to the Sort
MOV
JSR

TAG: .BYTE
.WORD
.WORD
.WORD

#TAG,RS
PC,RELES

3,0 ;3 ARGUMENTS TO RELES
<ADDRESS OF ERROR CODE>
<ADDRESS OF THE SIZE OF THE RECORD, IN BYTES>
<LOCATION OF THE FIRST WORD OF THE RECORD~
5-7

INTERNAL OPERATION
For FORTRAN, the call is as follows:
CALL RELES(IERROR,IRECSZ,RECADR)
where:
IRECSZ

is the decimal byte count of the size of the record to
be released.
IRECSZ must be positive and less than or
equal to MAXREC (error 5 if not) (See Section 5.4.1.1).

RECADR

is the address
word-bound).

5.4.1.3

the

record

(error

not

MERGE - Merging the Scratch Files

This is called after all input has been read.
MOV
JSR

TAG: .BYTE
.WORD

#TAG,R5
PC,MERGE

1,0 ;l ARGUMENT TO MERGE
<ADDRESS OF ERROR CODE>

For FORTRAN the call is:
CALL MERGE(IERROR)

5.4.1.4

RETRN - Requesting an Output Record
MOV
JSR

#TAG,R5
PC,RETRN

TAG: .BYTE 3(4),0; 3 (OPTIONALLY 4) ARGUMENTS TO RETURN
.WORD <ADDRESS OF ERROR CODE>
.WORD <ADDRESS OF WORD TO CONTAIN RECORD SIZE>
.WORD <LOCATION OF FIRST WORD OF THE RECORD>
---------- OPTIONALLY ---------.WORD <ADDRESS OF WORD TO CONTAIN LOCATION OF RECORD>
For FORTRAN, the call is as follows:
CALL RETRN(IERROR,IRECSZ,RECADR[,LOCREC])
where:
IRECSZ

is the decimal byte count of the size of
record.

RECADR

is the address of buffer to receive record (error 15 if
not word-bound).

LOCREC

if RECADR is O, RETRN places the address of the record
in the location rather than actually moving the record.

5-8

the

returned

INTERNAL OPERATION

5.4.1.5

ENDS - Ending the Sort

MOV
JSR

TAG,R5
PC,ENDS

TAG: .BYTE 1,0 ;l ARGUMENT TO ENDS
.WORD <ADDRESS OF ERROR CODE>
For FORTRAN the call is:
CALL ENDS(IERROR)

5.4.2

Setting up the Keys

Before calling RELES, you must first set up the key area, usually
outside the record itself. This area must be contiguous and an even
number of bytes long. Round up if odd.
If you are using byte data, the bytes within each word of the key must
be reversed in order before the key is stored. If your key is an odd
number of bytes, a padding byte of Os must be added to your key before
the bytes of the final word are reversed. Thus, a key that looks like
this in the record:
first word

looks like this when it is stored:
first word

second word
A

("paddi~g"

byte) I

If you are using word data, you must move the data word by word into
the key area, starting at the highest key location and working
downward.
The address passed to RSORT in the fourth parameter is the address of
the most major WORD. SORT always does word compares, starting at the
word in the highest memory location, and, if necessary, working down
to the least significant byte in the lowest memory location.
The address to be used for KEYADR can be figured as follows:
assume
KEYSTR is the first word of the key area which is KEYSIZ bytes long.
Then KEYADR = KEYSTR+KEYSIZ-2.
NOTE
The comparison is logical, so all 16
bits of a word are significant. There
is no implied sign. The control program
must therefore organize the key data
passed to SORTS in a form that ensures
the
correct
sequence, even if the
original data may be algebraic.
5-9

INTERNAL OPERATION
Calculating the Size of the Work Area

5.4.3

The size of the work
calculated as follows.
IWKSIZ =!FILES

area,

IWKSIZ,

must

certain

minimum

* (512. * BIGBUF + (100 + MAXREC + KEYSIZ + 10.)}

If you supply space below this minimum, SORTS will keep decreasing the
number of files until either the above equation is satisfied or the
number of files drops below 3, which will then cause RSORT to return
with error code 17.
Any extra core will be used to expand the in-core sort area.
general, the more space supplied, the faster the sort.

Thus, in

Typical Calling Sequences

5.4.4

To sort the file IN.TXT to produce the file OUT.TXT:
1.

Open IN. TXT.

Call RSORT to initialize the sort.

Read the next logical record from IN.TXT.
go to Step 6.

Set up the keys from the new record.

Cail RELES to give the record to the sort, then loop back
Step 3.

Close IN.TXT.

Call MERGE to collate the data.

Open OUT.TXT.

Call RETRN to get the
records, go to Step 11.

10.

Write the record onto the output file;

11.

Close OUT.TXT.

12.

Call ENDS to clean up the sort scratch files.

sorted

If no

record.

data,

loop to Step 9.

NOTE
The routine RETRN indicates "no more
records" by returning a negative value
in the error code.

5.4.5

Linking SORTS Subroutines with Your Program

The actual sorting subroutines are contained in SORTS.OBJ
and
SIORMS.OBJ. You can link these to your own calling program using Task
Builder. For example, if your program name was PROG:
TKB PROG=PROG,SORTS,SIORMS[, [l,l]RMSLIB/LB]

INTRODUCTION TO THE APPENDICES

Appendix A lists the ASCII character set and a subset of the EBCDIC
code.
Table A-1 lists the ASCII code values in ascending octal
sequence with the corresponding EBCDIC values.
You will find this
useful for comparing values when requesting a modified ASCII sorting
sequence, or when defining forced key fields.
Table A-2 lists the
EBCDIC subset used when you request an EBCDIC sorting sequence. The
EBCDIC values are in ascending order with the corresponding ASCII
values.
Appendix B contains two tables to help you to determine the digit and
zone values of printable characters. For example, the characters B,K,
and S have an equal absolute digit value of 2~ however, the "eleven"
punch denotes a negative value. The digit values of J through R are
negative (-) 1 through 9.
Appendix C gives examples of SORT applications.
Appendix D discusses SORT internal operations.
for the benefit of the advanced user and
successful SORT use.
Appendix E guides your installation of SORT.
Appendix F lists RMS status codes.

This is information
is not essential for

APPENDIX A

CHARACTER SETS USED BY SORT

Table A-1:

The ASCII Character Set with Corresponding
Codes

Table A-2:

The Subset of the EBCDIC Character Set
SORT When EBCDIC Sorting is Requested

A-1

EBCDIC

Used

CHARACTER SETS USED BY SORT
Table A-1
The ASCII Character Set with Corresponding EBCDIC Codes

__ .,_·-···--··-·--

Character

ASCII
Code

---·-·------·-·
EBCDIC
Code

NUL
SOH
STX
ETX
EOT
ENQ
ACK
BEL

000
001
002
003
004
005
006
007

000
001
002
003
067
055
056
057

BS
HT
LF

010
Oll
012
013
014
015
016
017

026
005
045
013
014
015
016
017

OLE
DCl
DC2
DC3
DC4
NAK
SYN
ETB

020
021
022
023
024
025
026
027

020
021
022
023
074
075
062
046

CAN
EM
SUB
ESC
FS
GS
RS

030
031
032
033
034
035
036
037

030
031
077
047
034
035
036
037

04 0
041
042
043
044
045
046
047

100
ll 7
177
173
133
154
120
175

050
051
052
053
054
055
056
057

115
135
134
ll6
153
140
ll3
141

0
1
2
3
4
5
6
7

060
061
062
063
064
065
066
067

360
361
362
363
364
365
366
367

8
9

070
071
072
073
074
075
076
077

370
371
172
136
114
176
156
157

. ··-

_...,_

--~~----

FF
CR

us
SPC

#
$
%
&

>
?

-··-·-····-···-~--~

EBCDIC
Code
-------- 1 - - - - - - - - - jr------ ______,
@
100
174
301
A
101
302
B
102
303
103
c
304
104
D
305
105
E
106
306
F
307
G
107

Character
-···

H
I
J

K
L
M
N
0
p

Q
R

s
T

v
w
x
y

z
[

a
b

c
d
e
f
g

h
i
j
k

1
m
n
0

p
q
r
s
t
u
v
w

x
y
z

DEL

-~·-·--~-~-~--~~~~-

A-2

ASCII
Code

llO
lll
ll2
ll3
ll4
ll5
116
ll 7

310
3ll
321
322
323
3 24
325
326

120
121
122
123
124
125
126
127

327
330
331
342
343
344
345
346

130
131
132
133
134
135
136
137

347
350
351
ll2
340
132
137
155

14 0
141
142
143
144
145
146
147

171
201
202
203
204
205
206
207

150
151
152
153
154
155
156
157

210
211
221
222
223
224
225
226

160
161
162
163
164
165
166
167

227
230
231
242
243
244
245
246

170
171
172
173
174
175
176
177

247
250
251
3 00
152
320
241
007

CHARACTER SETS USED BY SORT
Table A-2
The Subset of the EBCDIC Character Set used by SORT
when EBCDIC Sorting is Requested 1

Character

EBCDIC
Code

ASCII
Code

Character

·-----

NUL
SOH
STX
ETX
HT
DEL

·--1~-

203
204
205
206
207
210
2ll
221

143
144
145
146
147
150
151
152

p
q
r

222
223
224
225
226
227
230
231

153
154
155
156
157
160
161
162

s
t
u
v
w
x
y

241
242
24 3
244
245
246
247
250

176
163
164
165
166
167
170
171

251
300
301
302
3 03
3 04
305
306

17 2
173
101
102
103
104
105
106

307
310
3ll
320
3 21
322
323
324

107
llO
lll
175
ll2
ll3
ll4
ll5

325
3 26
3 27
330
331
340
34 2
343
344
345

ll6
ll7
120
121
122
134
123
124
125
126

346
347
350
351
360
361

127
130
131
132
060
061

362
363
364
365
366
367
370
371

062
063
064
065
066
067
070
071

000
001
002
003
Oll
177
013
014

c
d
e
f
g
h
i

015
016
017
020
021
022
023
026

015
016
017
020
021
022
023
010

LF
ETB

030
031
034
035
036
037
045
046

030
031
034
035
036
037
012
027

ESC
ENQ
ACK
BEL
SYN
EQT
DC4
NAK

047
055
056
057
062
067
074
075

033
005
006
007
026
004
024
025

SUB
SPC

077
100
ll2
ll3
ll4
ll5
ll6
ll7

032
040
133
056
074
050
053
041

120
132
133
134
135
136
137
140
141
152

046
135
044
052
051
073
136
055
057
174

153
154
155
156
157
171

054
045
137
076
077
140

172
173
174
175
176
177
201
202

072
043
100
047
075
042
141
142

FF
CR

SI
OLE
DCl
DC2
DC3
BS
CAN
EM
FS
GS
RS

[

<
(

+
!
&

l
$

*)
;
t or

,
%

or
>
?

#
@
I

=
II

a
b

ASCII
Code
-

000
001
002
003
005
007
013
014

EBCDIC
Code

1
m
n
0

z
A
B

D
E
F
G
H
I
J

K
L
M
N
0
p

Q
R
\

w
x
y

z
0
1

2
3
4
5
6
7
8
9

1 This table shows only those EBCDIC characters for which there are
ASCII equivalents. It therefore illustrates the collating sequence
SORT applies when the Header Specification contains E in column 26.

A-3

APPENDIX B
PRINTABLE CHARACTERS

Table B-1:

Printable Characters Grouped by Equal Digits

Table B-2:

Printable Characters Grouped by Equal Zones

B-1

PRINTABLE CHARACTERS

Table B-1
Printable Characters Grouped by Equal Digits

Group
.,.___-----+·

Character
blank

DEC 029
Card Code

12-5

11-5

0-5

12-6

No punches

12-1

11-1

11-6

0-1

0-6

r--·----- . ···-······-+······

Character

!-------·-· .. ---·--+-- . -·····-----··----------·--· ---------- ·-------

Group

..... ····-·-··- ----------------·-······---+ ·-----·-···----·-·

DEC 029
Card Code

- -1--

--·· ········--······---- ------+------------+--·-·-------- ------~

12-2

12-7
11-7

11-2

0-2

0-7

12-8

·- '·-···-········------

--------------+-------- - - -------------------+------~

12-3

11-3

11-8

0-3

0-8

12-9

11-9
0-9

-----~

------------··-·----+------------- -----··+------

12- 4

11-4

0-4

9
..

__________,______________-"-----····-··-·--

B-2

·--

9
···-·-------'-----------

PRINTABLE CHARACTERS

Table B-2
Printable Characters Grouped by Equal Zones

---- ---- -___,..,........----- --- .. -

--·-------·-

Group

Character

DEC 029
Card Code

Group

(

'"-~

1------

-----~

--------- - - - - - - · -

12-5-8

0-4-8

12-6-8

0-3-8

12-3-8

0-1

12-4-8

0-6-8

12-7-8

0-7-8

12-1

0-2

12-2

0-3

12-3

0-4

12-4

u
v

12-5

0-6

12-6

0-7

12-7

0-8

12-8

0-9

12-9

DEC 029
Card Code

Character
?-

, _,,

0-5

---

···· t--------··

-----r---

Blank

No punches

11-2-8

7-8

11-3-8

3-8

)

11-5-8

5-8

11-4-8

2-8

;

11-6-8

6-8

11-1

4-8

11-2

11-3

11-4

11-5

11-6

11-7

11-8

11-9

--·- --

B-3

APPENDIX C
SORT PROGRAM EXAMPLES

EXAMPLE 1:

SALES LIST

Input Record Positions
1-5
6-39
52-59
60-65
66-70
71-77
78-80

Customer Number (CUSTNO)
Customer Name (CUSTNM)
Date
Item Number (ITEMNO)
Item Quantity (ITEMQY)
Price
Blanks

This Record Sort is primarily intended to produce reformatted records.
The structure of the input file is indicated above.
The input file contains only one record type, therefore no Record Type
Specifications are needed. The keys for the sort are CUSTNO, ITEMNO,
and ITEMQY, with the last key sorted in descending sequence.
The
output file will have records ordered from lowest to highest customer
number, with the lowest to highest item number records within each
customer number category, and the highest to lowest item quantity in
each item number category.
The order of the data field specifications reformat the output record.
Note the use of the blank in columns 78 and 79; the data will be set
off in field columns.

c-1

SORT PROGRAM EXAMPLES
1 2

SORT SPECIFICATIONS

Page~

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

75 76 77 78 79 80

~:~;~f~otion lsl"li.IEkil I

--····--·------·····-·····-··········-··-·-····----------------------SORTR,T

Mode of Processing
0

SOR TR
SOR TT
SORTA
SORT!

Line

Total
Length of
Key Fields

Output
Record
Length

NOT USED

Comments
(Program Identification)

Forced
Lme

LL~~
!!. z 0

Field Location

I-----,

NOT USED

~ ~ ~ 1-----.,.-----1~

From

Field Name

Comments

I
I

~~.i....4-4-+-'~''~'~'-"+'~'-"-~'~'~16~17.+-~i-+:'c...'.:.:_21:~2 2_~ !.~ ~ -~~ ~!...~~ 29 30 31 32 3l 34 J5 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 5J 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 12 73 74

c1'1srr lol
1
"'

+···

I··

.. ,

_:!ljjelr11 C'([o

-1-l

1.1-1·~ Tl~'f

cul's'rto1lni

+···

I t-t

-f-

-+·

I·

t-+l-~

..... - f .
j..

·l··

I··

··-

I-

~--~

+····-t-1

Page~

FIELD SPECIFICATION

.+
t·

t- -

75 78 77 78 79 80

~;:~;~f~otion 1~11\l)..,l«fal I

Forced

I-----,

NOT USED

Field Name

Comments

I
I

~~O· ~.! ~~ ?.3 ~~ ~-~-·--~~.. ~7 __28 29 30 JI 12 ll J4 J~ J6 37 38 JIJ 40 41 42 43 44 45 46 ~7 48 49 50 51 52 SJ 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74

t+
I-

i··•·

-·

I-•

+··
- ~ --1

+ - t-t-+-+·

I-+-

+· ··t·

....

I ·- -

1 3

1 4

Figure C-1
EXAMPLE 2:

Sales List

OVERTIME ANALYSIS

Input Record Positions
1-3
4-8
9-39
40-43
50-53
80

Department Number (DEPTNO)
Employee Number (EMPNUM)
Employee Name (EMPNAM)
Standard Number of Working Hours (STDHRS)
Total Number of Hours Worked (HOURS)
Plant Identification Code

This Record Sort produces a record of employees by department number
in the main plant (M) who incurred overtime.
The Record Type
Specification shows the use of both an Include and an Omit line; with
a field and a constant Factor 2 comparison.

C-2

SORT PROGRAM EXAMPLES

"D"DD"D
~ ~
~ ~~~~~S~A~~~<;!~
E Q u I p M E N T

SORT SPECIFICATIONS

Page~

Dete _ _ _ __

HEADER SPECIFICATION

75 76 77 78 79 80

~:~;~f~ation lo&ll=\INl~li.I

Programmer

RECORD TYPE SPECIFICATION
Factor 1

0
Line

~ ~--------1 NE

::: .e a:

~ ~~

I-

Field Location

From

Factor 2

1-r-----------------1
f--- -

- Constant

I- Flel;Lo;;i.;
From

51¢

!ile>tLJ~lr1

!416

r- - - -

9 10 II 12 13 14 15 16 17 18 '.9 20 21 22 23 24 25 26 21'2s

~ 1 jq tl>I
~2-t-P=I jcj

~---1

Comments

2~ 30r2-l ~?~~6 37 38 39 40 41 42 43 44 45 46 47,48 49 50 51 52 53 54 55 56 57 ~. SJM!M3_,6_4 65 66 67 68 69 70 711:,1.?~

~- -1-++-i Ito v ~1~
111 .i~Ji
+ t..jo c.~T j.tjojNlcn Aji:jtl

jt+-J3

!S'jszsJeJ&ic "1 1 -t--i

o 3

-1

Record Name

t--t--1

-f--14- - t---+-t-1- f--t--1

e xc!c::.J£!!> ~IT"AN~ A~
P L. A t<~I o(NJi..)Y
~ t-

t·

0 4

ro 5
r-

·-·-

o 6

rt- -

FIELD SPECIFICATION
Forced

Line

t----,
Field Name
s

Comments

I
I

a 9 10 11 12 ll 14 15 16 11 1s 19 20 21 22 23 24 2s 26 21-~~ .. ~~o 2_1 !~~~JS 36 11 JS 39 40 41 42 43 44 45 46 47 48 49 so s1 s2 sJ 54 ss ss s1 sa 59_60 61 62 63 64 65 66 67 68 6!,~ ~-1~

NP'I

~1
._:i.1a

FPIJ>I

0
i- /

~1»1

~ 0 - FtJ>lq
i--1- 1 F~P>I

1
5¢
I

!3j ·

~ e P Ii n o

5!~

~lo(vRF>J

jlt

]8j

l~~

IE!tnl".~Jv fl>j

~I~' f( AMJ -

1--1

1---i

t--+-i -

-- -+

1 2
i--1 3

r---

I-

1 4

Figure C-2
EXAMPLE 3:

Overtime Analysis

EMPLOYEE LIST

This Tag Sort illustrates the use of forced key field specifications.
The sort is based on the plant identification code in column 80 of the
input records. In order to sort them in the order M, W, P, the M in
column 80 is forced to a 1, the W is forced to a 2, and the P is
forced to a 3. The output record entry is unchanged.

C-3

SORT PROGRAM EXAMPLES

11n•nn•nEQU
PM ENT
~ ~
~ ~~~~~s~.~~uls~':!
I

SORT SPECIFICATIONS

75 78 77 78 79 80

Page~

Date _ _ _ __

HEADER SPECIFICATION

~:~;~,~ation lc::lmlP l1-ls !fl

Programmer - - - - - -

SORTR,T

Mode of Processing

Line

SOR TR
SOR TT
SORTA
SORTI

Total

Length of
Kev Fields

Output

NOT USED

Comments

Record

(Program Identification)

Length

RECORD TYPE SPECIFICATION

0
Line

Factor 1

~ lLI------~

'; ~ a~
o

? u.~ §

. --···-··-····-..-·---··
Factor 2

---·---·~-

1-r---------------1

. . . . . . . . .:··7· .. l1
~~

Field Location

From

LE ~

i---------t- -

- Constant--------

- I

Field Location

From

I- -

Record Name

ff·

Comments

6 37

l-+-1
I

a 2

a 3
1---+-

o 4
1-+-+-io 5

l-+--1-0 6

~ ~al.,· . ··~··~ . ,.i................................. ·~n·_._[j_

vrr-1

. ·- -- -- -

_.. --

.L....J_,___ L....J________

-·

FIELD SPECIFICATION
r--"T""'"lr-r-.-------~-..,.--------------

Forced

----,

NOT USED

Field Name

Comments

-··
I·"

I I-

-1-··

+-·

t:\l~~ ~Ii t>~lt' r.i(;, Aj"-Jh Y
_
w~p~~ ~Ecpa~ oµ~~~tli

1 2
~-+-

1 3

l-+-

1--1 ·

1--1-+-·

1 4

Figure C-3
EXAMPLE 4:

Employee List

BONUS ANALYSIS

This Index Sort illustrates the use of multiple Include and Omit
lines.
The first category of Omit lines excludes all main plant
industrial employees' records that do not show bonus eligibility; the
second set of Omit lines excludes any employees' records from any
other plant that do not show bonus eligibility. The line with an I in
column 6 and no other entries includes all other records; i.e., the
records of industrial staff employees from the main plant and all
employees from any other plant who are eligible for a bonus.

C-4

SORT PROGRAM EXAMPLES
1

SORT SPECIFICATIONS

75 76 77 78 79 80

Page~

Date _ _ __

HEADER SPECIFICATION

i:~;~f~ation !iJO '"'Ill{ IA Id

Line

SOR TR
SOR TT
SORTA
SORTI

Total

Length of
Key Fields

NOT USED

f
RECORD TYPE SPECIFICATION

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

~ "-

Line

:o~: g~c !:!~iii
8u

~2 oA~j

~IAl<::i

~14

~ 1~A~I
~

0 6

1---

LT
LE ~

-- C o n s t a n t - - - f- - - - - I
Field Location
j
From
To
I

------

I- -

Record Name

Comments

9 20 2122 23 24 2s 26 21'2s 29 JO JI J2 JJ 34 35 36 11 JS 39 40 41 42 43 44 45 46 47148 49 so 51 s2 53 54 ss 56 57 sa 59 60 61 62 63

s~ejQjcN

5j~~~~

7~ Nle;c 1

j~~

I~~

-1 t--+--'-oc.Apj:tjo[M_ i=-1 l"'f:t\.tj_.Z f'LAlll[i1

53~y~

riov~s

6\~69 71!~!~ 73r~

- ~usl• ~xq~~~ ~jrA~~~I~
o_ _ ~lc.iftl"ll
_ _ _ El6.['4 - :t!N\i>Ju~Jtj~:rjA '- ~ji"iAjf1Fj ~!Mla..j~

Factor 2

g~
Field Location
From
To

a 9 10 11 12 13 14 15 16 1118

~- 0 lq
: :1

Factor 1

IEJY51,_

t.--1

i...lqc~llj.J:.jojrl

Ho1J~s

- A~jov~ 601~~s ~l~1l~ o~\~V

FIELD SPECIFICATION

------·- --------

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

.......,

Forced

.._ ~ ~

Line

8_2~

~
S

! ~ From
7

Fit~

0 8
1--1
0 9

Fr{[c
F MIC

1 0

1_F_

g~~

o 7

1---1---

Field Location

rllD

1Jo""

~---.......,

NOTUSED

~ ~

E c

Comments

Field Name

~ 8 8

I --,- -,--,--~ -- -- .,4

9 IO 11121314151617181920212223242S 26 2128293031323334 3536J7J839 14041424344454647484950Sl525354555657 5859606162636465666768697011727374

I~~

- ~~I~!~~

~l~

;lq

lit

--1

-r- ----

cAT"&-<:r'/

sj¢ .

1-Joc.AiTIN

e-le11'1I ~vl1t1

f--

1 1

I-+-,

·--

1--1 3
1---

t--1-i

- -+-

1 4

Bonus Analysis

Figure C-4
EXAMPLE 5:
1-2
3-10

50-55
70-75

STOCK ORDER LIST
Warehouse Identification Code
Stock Number
Quantity of Items Required to be in Stock
Quantity of Items in Stock

This is an ADDROUT Sort intended to produce a file of stock numbers by
warehouse. Those records whose stock quantity level is lower than the
accepted level are listed first in each warehouse category.
A
priority number (1 for under-stocked items; 2 for items of sufficient
quantity) is forced into position 3 of the output record.
The output file records will be in the following order:
Warehouse 1

Stock Numbers of Items whose qua~tity is low
Stock Numbers
acceptable

Warehouse 2

Items

whose

quantity

Stock Numbers of Items whose quantity is low
Stock Numbers
acceptable

C-5

Items

whose

quantity

SORT PROGRAM EXAMPLES
etc.

Warehouse 3

SORT SPECIFICATIONS

75 76 77 78 79 80

Page~

Program

Identification

lsfT~~
~Y

HEADER SPECIFICATION
Programmer ~-----

.----.-..------....----.-.------........T""""1r---·---.--·-----------··-·-~

Mode of Processing

0
;;c

SORTA
SOR TT
SORTA
SORT/

Line

Total
Length of
Key Fields

SORTR,T

-iii ax

NOT USED

£~ ~

Comments

Output

8 b ~ ~:~;;~
- z 9'

;j'

• • 10 II 12 13" 15 16 17" "20,, 22",.,, ,. ,, ,, ,.

!Program Identification)

!: ;;f.~r~~IYhl.5EirI'Liriri.iirII 6. 65 66 67 68 69 70" 72 73,.

RECORD TYPE SPECIFICATION

---------- -- -----------

------- ---------

I- -

Comments

-------···-·------------- -----------------------

Factor 1

g~~

Lone

I!. ~ fl
.
? ~ s~c~---

8
7

From
8

Factor 2

~~ f-

- Constant - - - _______........

I - -:- -

GE
LT
LE ~

- - I
Field Locat1?~ __J

From

Record Name

Jt QoA•~J*Y [~S~ ~~·, ~Ari[:•1 ~~£V~~~ ~
I

10~ 14 IS~ 1718~__!20 :1'122r~24 2!1 16 2Jl<'_819 JO 3_:1 ~2 JJ 34 35 36 37 38 l! ~i:,.~-~~ 45 46 ~714_8_ 49 50 51 52 53 5! 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74

~ ~ 7~ r~~ 5~1 J]J ,,~

FIELD SPECIFICATION

-----------------------

- -·r----··-··1 -·
Forced

f-----,

NOT USED

Field N"me

Comments

0 8
0

. l---+-+-

f--

1-+--+- -+-

1 2
1 3

-+-

Figure C-Sa

Stock Order List, Page 1

C-6

SORT PROGRAM EXAMPLES
1
Page

HEADER SPECIFICATION

[[J

Mode of Processing

SORTA
SOATT
SORTA
SOATI

Lme

Total

Length of
Key Fields

Factor 1
EQ
0
~ u . 1 - - - - - - - - I NE

~ u- ~ ojN_u~
~

_ _

Field Location

From

Factor 2

1-------------------1

t ~ __ =~ons1ant
Field Location

From

----

__ _..,. .....__ _ _ _ _ I

Comments

Record Name

r··~~l~···::····:.i:;;t:If ··.J~'"I11fill·r·· . ·~:. . . . , . . . . . . . . . . . u~· :~~·· ~
..

FIELD SPECIFICATION

! ~~
~ u

Field Location

From

f-----r---li
To

QEd

I-----,
F1ekl Name

8 8

Comments

I
I

9 10 111213141516171819202122232425 :>6 2728i.'93031323334 3536313839404147434445464748495051525354!>55657 5859606162636465666768697071727374

0 8

1"'1

32
I If

f
1 2

Figure C-5b

Stock Order List, Page 2

C-7

APPENDIX D
INTERNAL OPERATIONS

D.l

PROCESS DESCRIPTION

The material in this Appendix is included for the benefit of the more
experienced programmer.
This information is not necessary for the
successful use of SORT by the less experienced user.
The SORTS subroutines, which make up the bulk of the SORT program,
consist of two phases: a sorting and distribution phase and a merge
phase.
During the sorting and distribution phase,
the
RELES
subroutine accepts records from the input file and places them in an
internal sort area called the sort tree.
Each time the tree is
filled, RELES distributes intermediate strings or runs of sorted
records to the scratch files.
The average length of these runs
depends on the size of the record and its key. Also, the more nearly
sorted the input file is, the longer the runs are.
Because the
algorithm of the SORT package is replacement selection, the average
length of a run is twice the number of records that fit into the sort
area.
At the end of the input file, the RELES routine finishes the sorting
and distribution phase and RSORT calls the merge phase. The merge
phase consists of two subphases: an intermediate merge, the MERGE
subroutine, and a final merge, the RETRN subroutine. The MERGE
subroutine reads the series of runs from the scratch files, merges
them into smaller numbers of longer runs and writes them out to the
scratch files until there is only one run per scratch file. The RETRN
subroutine then merges the runs together to create a single stream of
records to be returned to you one at a time.

D.2

THE SORT WORK AREA

The entire work area of SORT is in the psect AAAAAA. Psect AAAAAA has
an initial length of O, which you can expand by using the EXTSCT
command in the task build command file.
The size of this psect
controls the ultimate task image size. In general, the larger the
AAAAAA psect is, the faster SORT runs.
SORT calculates the size of the work area by subtracting the address
of the global location WA from that of the global location BOTTOM.
These two global symbols mark the upper and the lower limits,
respectively, of the SORT work area.
The SORT work area is used for several purposes, which are illustrated
in Figure D-1. Figure D-la shows the layout of the work area at the
end of the command decoding phase. RMS uses space near the lower end
of the work area for file control blocks and SORT uses space near the
upper end for internal key information. The upper area is initially
D-1

INTERNAL OPERATIONS

set up to hold 10 keys. Its size is stored in the global location
KYAREA in the segment SORTCX. To change the size of the key area,
enter the following command in the task build command file:
GBLPAT=SORTCX:KYAREA:x
where:
x

is the number of keys times 30 and is expressed in octal.

The pointers WORKHI and WORKLO contain the addresses of the upper and
lower limits, respectively, of the remaining work area space and are
maintained throughout the sorting process.
Figure D-lb shows the physical layout of the SORT work area before
RSORT begins.
An additional area from the low end of the available
work area space is turned over to RMS for internal control blocks.
SORTS also sets up the input file buffer. Note the changed positions
of WORKHI and WORKLO. The positions have changed because the values
of the addresses in these two locations have changed to reflect the
new boundaries of available work area space.
If the input file is designated as sequential, SORT uses the RMS large
disk blocking feature. Large disk blocking, also called big blocking,
means that if you have enough core, you can call several blocks of
your input file off the disk at once, instead of just one block at a
time. Big blocking can be applied to all the files you use in your
sort and, if you can use it, speeds up the sorting process by cutting
down the number of times you have to do a disk seek.
The SORT control program initializes the
sorting process as follows:

work

area

for

the

actual

RSORT divides the available work area core space into
quarters.
One quarter, rounded down to the nearest 512-byte
block, is used as an input block buffer.

The number of blocks in the input block buffer is passed as a
parameter to RSORT to control its scratch file blocking
factor.
To override the scratch file blocking factor,
specify an explicit blocking factor with the BUCKETSIZE
switch on the input side.

The amount of space remaining in the work area after the
space for the input block buffer is deducted is passed to
RSORT as a parameter.

If the big blocking factor is 0, then half of the space left after the
input block buffer is deducted is allocated for scratch file buffer
space. If the big blocking factor is not O, then that many 512-byte
blocks are allocated for scratch file buffer space.
(The big blocking
factor is an octal number.) In either case, the leftover space is used
to store the sort tree and the record areas. You can change the big
blocking factor on the scratch files by using this command at task
build time:
GBLPAT=SORTS:$RBGBF:x
where:
x

is the big blocking factor in octal.

D-2

INTERNAL OPERATIONS
BOTTOM-r-------,----RMS INTERNAL
CONTROL BLOCKS
WORKLo-----------1

SCRATCH FILE
BUFFER

FILE 1
BUFFER

FILE 2
BUFFER

SORT TREE

FILE 3
BUFFER

FILE 4
BUFFER
WORKHI - - - - - - - - - · - ;

INPUT FILE
BUFFER

FILE 5
BUFFER

WORKHI - - - - - - - - - - i
KEY AREA
WA---..L------J~----

a. AFTER COMMAND
DECODING

b. BEFORE THE SORT

Figure D-1

c. DURING THE SORT

d. AFTER MERGE

SORT Work Area Allocation

The first record in the input file is read and the actual sorting
process begins. SORTS handles the actual sorting and distribution of
the input records. The physical core layout for the SORT work area
resembles Figure D-lc.
At the end of the record distribution process, the input file buffer
is no longer needed, so SORTS releases it. If the output file does
not need more than two blocks of buffer space - that includes all
sequential files and any relative files with bucketsizes less than
three - the input buffer space is added back to the work area and
WORKHI is adjusted accordingly.
At the start of the merge phase, SORTS divides the available work area
space into as many buffers as there are scratch files specified.
Figure D-ld shows the default - five buffers. One of the buffers is
larger than the rest and the raiio between its size and the size of a
small buffer is:

~:l
where:
n

is the number of scratch files.

You can see that the more scratch files you specify, the larger the
discrepancy in sizes between the large buffer and one of the small
ones.
After the intermediate merge subphase and before the final merge,
SORTS turns over the large buffer to RMS for use on the output file.
Big blocking is enabled for sequential files, and the final merge
takes place.

D-3

INTERNAL OPERATIONS
D.3

PERFORMANCE PARAMETERS

Two main concepts you must consider
efficiency of SORT are:

when

you

try

increase

•

Determining the most efficient size of the work area

•

Dividing the work area size among competing needs

The advantages of giving more space to the input and
at the expense of the sort tree are:
•

Larger blocking factor

•

Lower disk access and seek time

output

the

buffers,

But a smaller sort tree results in:
•

More intermediate runs

•

A longer merge phase run

The default proportions for the input and output buffers and the
tree are 25, 25, and 50 percent, respectively.

sort

During the merge phase, RETRN uses almost all of the work area space
for scratch file buffers.
By assigning a large number of scratch
files, you reduce the number of merge runs and general record
shuffling.
But if you lack the core space necessary for a large,
fast-running sort, more scratch files can be a handicap. More scratch
files in a small space means smaller files can be a handicap. More
scratch files in a small space means smaller I/O buffers per file and
therefore greater disk access and seek times.
Table D-1 shows you how to change seven factors that affect the speed
and efficiency of SORT. The Result column reflects the effect of an
increase in the factor being on the run time of the sort.
Plus (+)
means a faster sort;
minus, a slower one; zero, no effect. The
Change column shows the variables that you should change to affect the
factor being considered. The Time and Place column legends translate
as follows:
•

TKB - make the change to the Task
task build time.

•

RSORT - make the change to the RSORT parameter list.

•

IFS - make the change to the input file switch on the command
file at task build time.

D-4

Builder

command

file

INTERNAL OPERATIONS
Table D-1
User-Accessible SORT Parameters
..

Result

Time and
Place

EXTCST=AAAAAA:n
IWKSIZ

+
+

TKB
RS ORT

Input Buffer Size

/BU:n

IFS

Scratch File
Buffer Size

GBLPAT=SORTS:$RBGBF:n
BIGBUF

+
+

TKB
RS ORT

Key Area Size

GBLPAT=SORTCX:KYAREA:n

TKB

Number of
Scratch Files

GBLPAT=SORT:FILES:n

TKB

Scratch File
Clustersize

GBLPAT=SORTS:$RSCLV:n
SCRLV
/SI:n

+
+
+

TKB
RS ORT
IFS

Contiguous
Scratch Files

GBLPAT=SORTS:$RSCTG:l
/CO

+
+

TKB
IFS

Scratch File
Preallocation

GBPAT=SORTS:$RPALQ
PRIALQ
/AL:n

+
+
+

TKB
RS ORT
IFS

SORT Parameter

Variable to Change

Work Area Size

Detrimental if you lack sufficient core to run a large, fast sort.
NOTE
The /AL, /CO, and /SI switches apply only to the
temporary scratch files when used on the input
side. This usage is intended only for programmers
who are completely familiar with the effects of
these switches.

D-5

APPENDIX E

SORT INSTALLATION

E.l

GENERAL COMMENTS

SORT is an independent task that can be run under control of RSTS/E
(V06B-02), RSX-llM
(V3.0), or IAS (V2.0). SORT is neither shareable
nor reentrant, so multiple calls to SORT create multiple copies of
SORT in core storage.

E.2

SORT KIT CONTENTS

The SORT kit contains individual command files for you to use to build
SORT on your operating system. These files are:
•

For full RMS-llK (including indexed) support
SRTRIM.CMD - RSTS/E with RMSll Runtime System
SRTEIM.CMD - RSTS/E with RSX Runtime System
SRTMIM.CMD - RSX-llM and IAS

•

For sequential and relative only RMS-11 support
SRTRRM.CMD - RSTS/E with RMSll Runtime System
SRTERM.CMD - RSTS/E with RSX Runtime System
SRTMRM.CMD - RSX-11 M and IAS

When you build SORT you must use one of these files,
depending
which operating system and what kind of RMS-11 support you have.

The Task Builder command file contains a number of parameters that can
be changed and options that can be included by the system manager to
tune SORT to your installation's needs. These parameters control the
size of the in-core work area (the scratch pad) which controls the
task size, and the default logical unit and device assignments for the
files SORT uses.
Chapter 5 and Appendix D contain a full treatment of
these parameters and other options.
You can change the command file
after it has been taken from the distribution medium and before the
task build is started. To change just the task image size, see
Section F.3.
As supplied by DIGITAL, the command file builds the largest possible
task
(28K words on RSTS/E, 32K words on RSX or IAS) on account number
[l,2], the library account, under the name SORT.TSK. The program uses
five scratch files as the default and finds and puts all files on the
public disk structure.

E-1

SORT INSTALLATION
E.2.1

Building the SORT Task for RSX-llM or IAS

RSX-llM users can build and install SORT by using this procedure:
1.

Use FILEX or PIP to transfer SORT from the distribution
medium to a privileged account on public disk space:
>FLX SY:/RS=MMO: [1,2]*.*/DO

At this point make any changes to the
file:

Build the task:
>TKB @SRTMIM

Builder

command

(for full RMS-llK
support on RSX-llM or IAS)

Install the task image:
>INS [l,2]SORT

Task

(the default task
name is •.. SRT)

Erase the files if no longer neededi
>PIP @SRTERS

Invoke SORT:
>SRT
SRT>

E.2.2

(SORT is now ready)

Building the SORT Task for RSTS/E
NOTE
The following procedures for installing SORT-11 apply only to
RSTS/E V6C and earlier versions.
For instructions on SORT-11
installation, refer to the RSTS/E System Generation Manual.

RSTS/E users can build and install SORT by using this procedure:
1.

Use PIP to transfer the file from the distribution medium
a privileged account on public disk space:

PIP *.*/CO=MMO:[l,2]*.*
READY
2.

At this point make any changes to the
file.

Build the task:
RUN $TKB.TSK
TKB> @SRTRIM
TKB>//
READY

Task

Builder

(full RMS-11 support)

E-2

command

SORT INSTALLATION

Make the task image executable by all users:
PIP $SORT.TSK<l04>/RE
READY

Add the "SRT" CCL command:
RUN $UTILTY
UTILTY V06B-03 RSTS V06B-02 TS/l
#CCL SRT-=$SORT.TSK;O

READY
6.

Spool off and save the build map:
QUE SORT.MAP/DE
READY

Erase the files if no longer needed:
PIP @SRTERS
READY

E.3

CHANGING THE SORT TASK IMAGE SIZE

The total size of the installed task image is controlled by the size
of the psect AAAAAA, the scratch pad work area. To make the task
smaller, cut down the size of the psect as follows:
1.

Use an editor to access the proper command file.

Find the line "EXTCST=AAAAAA:XXXXXX".

Copy the number represented by XXXXXX here.

Determine the desired task size reduction
in 4K word increments (e.g. 12KW = 3)

Multiply the number on line B by 20000(8).

Subtract line C from A using octal arithmetic.

Replace XXXXXX in the command file with the
value on line D.

Continue with the build.

E-3

APPENDIX F
RMS I/0 STATUS CODES

This appendix describes I/O status codes that can be returned to
program by RMS.

your

For some error conditions, RMS uses the RMS-status-code field of the
SORT error message to communicate additional information. Table F-1
lists the RMS status codes and what they mean.
Table F-1
RMS Error Status Codes
Decimal
Value

Description

-16

Operation aborted: out-of-stack save area or
structures corrupted.

-32

Files-11 ACP could not access the file.

-48

File activity precludes action (e.g., attempting to close
file with outstanding asynchronous record operation).

-64

Bad area identification number (AID} field in allocation XAB
(i.e., out of sequence}.

-80

Illegal value in alignment
allocation XAB.

-96

(or
Value in allocation quantity (ALQ} field in FAB
allocation XAB} exceeds maximum or, during an explicit
$EXTEND operation, equals O.

-112

Records in a file on ANSI-labeled magnetic tape are variable
length but not in ANSI D format.

-128

Illegal value in
allocation XAB.

-144

Invalid operation at AST level:
attempting to issue a
synchronous operation from an asynchronous record operation
completion routine.

-160

Read error on file header attributes.

-176

Write error on file header attributes.

boundary

allocation

type

options

in-core

(ALN}

(AOP}

field

data

(Continued on next page}
F-1

RMS I/0 STATUS CODES

Table F-1 (Cont.)
RMS Error Status Codes
--

Decimal
Value

---···---··---------------·

Description

-192

Bucket size (BKS) field
maximum.

FAB

contains

-208

Bucket size (BKZ) field in
exceeding maximum.

-224

Block length (BLN) field in a FAB or RAB is incorrect.

-232

Beginning of file detected on $SPACE operation
tape file.

-240

Private buffer pool address not a double word boundary.

-256

Private buffer pool size not a multiple of 4.

-272

Internal error detected in RMS-11;
contact a Software Specialist.

-288

Cannot connect RAB (i.e., only one
permitted for sequential files).

record

-304

$UPDATE attempting to change a key field that does not
the change attribute.

have

-320

Index file bucket check-byte mismatch. The bucket has
corrupted. No recovery possible for the bucket.

been

-336

Close function failed (RSTS/E operating system only).

-352

Invalid COD field in XAB or this XAB type is illegal for the
organization or operation.

-368

Files-11 ACP could not create file.

-384

No current record: operation not immediately preceded by
successful $GET or $FIND.

-400

Files-11 ACP deaccess error during $CLOSE.

-416

Invalid area number in DAN field of key definition XAB.

-432

Record accessed by RFA access mode has been deleted.

-448

1.
2.
3.

allocation

value

XAB

exceeding

contains

recovery

value

magnetic

possible;

access

stream

Syntax error in device name.
No such device.
Inappropriate device for operation (e.g., attempting
create an indexed file on magnetic tape).

-464

Syntax error in directory name.

-480

Dynamic memory exhausted:
insufficient
space pool or private buffer pool.

space

central

(Continued on next page)

F-2

RMS I/0 STATUS CODES

Table F-1 (Cont.)
RMS Error Status Codes
Description

Decimal
Value
-496

Directory not found.

-512

Device not ready.

-520

Device positioning error.

-544

Duplicate key detected, duplicates allowed attribute not set
for one or more key fields.

-560

Files-11 ACP enter function failed.

-576

Environment error;

-592

End of file.

-608

Expanded string area in NAM block too short.

-616

File expiration date not reached.

-624

File extend failure.

-640

Not a valid FAB:

-656

1.
2.
3.

operation not selected in ORG$ macro.

BID field does not contain FB$BID.

Record operation attempted was not declared in FAC field
of FAB at open time.
Invalid contents in FAC field.
FB$PUT not present in FAC for $CREATE operation.

-672

File already exists (attempted $CREATE operation).

-680

Invalid file ID.

-688

Invalid combination of values in FLG field of key definition
XAB (e.g., no duplicates and keys can change).

-704

File locked by another user -- your program cannot access
the file because its sharing specification cannot be met.

-720

Files-11 ACP Find function failed.

-736

File not found.

-752

Syntax error in file name.

-768

Invalid file options.

-784

Device full:

-800

Invalid area number in IAN field of key definition XAB.

-816

Index not initialized (this code can only occur in
field when STS contains ER$RNF).

cannot create or extend file.

the

STV

(Continued on next page)

F-3

RMS I/O STATUS CODES

Table F-1 (Cont.)
RMS Error Status Codes
Description

Decimal
Value
-832

Invalid IFI field in FAB.

-848

Maximum number (254) of key definition or allocation XABs
exceeded or multiple summary, protection, or date XABs
present during operation.

-864

$INIT or $INITIF macro call never issued.

-880

Illegal operation;
1.
2.
3.
4.

examples include:

Attempting a $TRUNCATE operation to a non-sequential
file.
Attempting an $ERASE or $EXTEND operation to a magnetic
tape file.
Issuing a block mode operation (e.g., $READ or $WRITE)
to a stream not connected for block operations.
Issuing a record operation (e.g., $GET, $PUT) to a
stream connected for block mode operations.

Illegal record
count field.

-912

Invalid internal stream identifier (ISI) field in RAB (field
may have been altered by user) or $CONNECT never issued for
stream.

-928

Key buffer address (KBF) field equals 0.

-944

Record identifier (i.e., the 4-byte location addressed by
KBF) for random operation to relative file is 0 or negative.

-960

Invalid key of reference (KRF) in RAB:
1.
2.

encountered

sequential

invalid

-896

file:

As input to random $GET or $FIND operation, or
As input to $CONNECT or $REWIND (in this case, ER$KRF is
returned for the first record operation following the
$CONNECT or $REWIND) •

-976

Key size equals 0 or too large (indexed file) or
to 4 (relative file).

not

equal

-992

Invalid area number in LAN field of key definition XAB.

-1008

Magnetic tape is not ANSI labeled.

-1024

Logical channel busy.

-1040

Invalid value in logical channel number (LCH) field of FAB.

-1048

Attempt to extend an area containing an unused extent.

-1056

Invalid value in LOC field of allocation XAB.
(Continued on next page)

F-4

RMS I/O STATUS CODES

Table F-1 (Cont.)
RMS Error Status Codes
Description

Decimal
Value
-1072

In-core data structures (e.g., I/O buffers) corrupted (this
code can only occur in the STV field when STS contains
ER$ABO).

-1088

Files-11 ACP could not mark file for deletion.

-1104

1.
2.

-1120

Maximum record number field contains a negative value
during $CREATE of relative file.
Record identifier (pointed to by KBF) for
random
operation to relative file exceeds maximum record number
specified when file created.

Maximum record size (MRS) field contains
operation and:
1.
2.

during

$CREATE

Record Format is fixed, or
File organization is relative.

-1136

Odd address in Name Block address
$OPEN, $CREATE, or $ERASE.

-1152

Not at end-of-file:
attempting a $PUT operation
sequential file when stream is not positioned to EOF.

-1168

Cannot allocate internal index descriptor:
insufficient
room in space pool while attempting to open an indexed file.

-1184

No primary key definition
indexed file.

-1200

Open function failed (RSTS/E operating system only).

-1216

XABs in chain not in correct order:
1.

XAB

(NAM)

present

field

during

Allocation or key definition XABs not in
densely ascending) order.
XAB of another type intervenes in key
allocation XAB sub-chain.

FAB

$CREATE

ascending

(or

definition

-1232

Invalid value in file organization (ORG) field of FAB.

-1248

Error in file prologue:
reconstructed.

-1264

Key position (POS) field in key definition
value exceeding maximum record size.

-1280

File header contains bad date
and
time
information
(retrieved by RMS-11 because a date and time XAB is present
during a $OPEN or $DISPLAY operation);
file may
be
corrupted.

file

corrupted
XAB

and

must

contains

(Continued on next page)

F-5

RMS I/O STATUS CODES

Table F-1 (Cont.)
RMS Error Status Codes
..--------~----------------

..... •••············ ·······-·--·--·, ~-,.·----·--·-··--·----------_..,

··----···-··

Description

Decimal
Value
f - - - - - - - - - t - - - - - · · · · · ·····-

-1296

Privilege violation:
operating system.

-1312

Not a valid RAB: BID field does not contain RB$BID.
to Section A.3 of this Appendix.

-1328

2.
-1344

1.
2.

access

the

file

denied

the
Ref er

Illegal values in record access mode (RAC) field of RAB.
Illogical value in RAC field (e.g., RB$KEY with a
sequential file).
Illegal values in record attributes (RAT) field of
during $CREATE.
Illogical combination of attributes (e.g., FB$CR
FB$FTN) in RAC field during $CREATE.

Record address (RBF) field in RAB contains
(block mode access only).

-1376

File read error.

-1392

Record already exists: during a $PUT operation in random
mode to a relative file, an existing record found in the
target record position.

-1408

Invalid RFA in RFA field of RAB during RFA access.

-1424

1.
2.

odd

and

-1360

Invalid record format in
$CREATE.
Specified
record
format
organization.

FAB

address

RFM

field

FAB

during

illegal

for

file

-1440

Target bucket locked by another task or
the same program.

another

-1456

Files-11 ACP Remove function failed.

-1472

Record identified by KBF/KSZ fields of RAB for random $GET
or $FIND operation does not exist in relative or indexed
file (for indexed files only, STV may contain ER$IDX).
Record may never have been written or may have been deleted.

-1488

$FREE operation issued but no bucket was locked by stream.

-1504

Record options (ROP) field contains
illogical combination of values.

-1520

Error while reading prologue.

-1536

Invalid PRV record encountered in indexed file;
corrupted.

-1552

Record stream active, i.e., in asynchronous environment,
attempting to issue a record operation to a stream that has
a request outstanding.

illegal

stream

values

file may be

--------·--·------------····--·---

(Continued on next page)
F-6

RMS I/O STATUS CODES

Table F-1 (Cont.)
RMS Error Status Codes
Description

Decimal
Value
I------+-----------------------

-1568

Record size specified in RSZ of RAB during $PUT
is invalid:
1.
2.
3.
4.

$UPDATE

RSZ equals O.
RSZ exceeds maximum record size (MRS) specified when
file created.
RSZ not equal to size of Current Record for $UPDATE
operation to a sequential file on disk.
RSZ does not equal MRS (for fixed format records).

-1584

Record too big for your buffer:
RMS-11 could not move
entire record retrieved by $GET operation to work area
(UBF/USZ). Note that this error does not destroy the
current context of the stream. Rather, the stream's context
is updated as if the operation had
been
completely
successful.

-1600

During $PUT operation, key of record to be written is not
equal to or greater than key of previous record (and RAC
field contains RB$SEQ) •

-1616

Illogical value in SHR field of FAB (e.g., FB$WRI
for sequential file).

specified

-1632

Invalid SIZ field in key definition XAB during
(e.g., specified size exceeds maximum record size).

$CREATE

-1648

During asynchronous record operation, RMS-11 has found that
the stack is too big to be saved (this code can only occur
in the STV field when STS contains ER$ABO).

-1664

System directive error.

-1680

Index tree error:

-1696

Syntax error in file type (e.g., more than three
specified).

-1712

Invalid address in UBF field of RAB:
1.
2.

indexed file is corrupted.
characters

UBF contains O, or
UBF not word aligned (for block mode access only).

-1728

Invalid USZ field in RAB (i.e., USZ contains 0).

-1744

Syntax error in file version number.

-1760

Invalid VOL field in allocation
contain 0).

-1776

File write error.

-1784

Device is write locked.

-1792

Error while writing prologue.

-1808

XAB field in FAB (or NXT
address .

______,____________________

.__

~~---

field

XAB

(i.e.,

XAB)

VOL

does

not

contains

odd

··--- - - - - - - - - - - - - - · - - - - - - - - - - - -

F-7

GLOSSARY

Alphanumeric Characters - the entire set of 128 ASCII characters.
ASCII Character Set - the set of 128 seven-bit American Standard
for Information Interchange characters (see Appendix A).

Code

Byte - the smallest addressable unit of information; eight bits.
Each ASCII character usually resides in a single byte.
Collating Sequence - the order into which characters are sorted
upon numeric values assigned to each.

based

Device - any of the peripherals referenced by the operating system.
Digit - in EBCDIC representation, the numeric value of the
four bits of a byte.

low-order

EBCDIC - an alternate character set used widely in the computing
industry (see Table A-2 for comparison with ASCII characters).
Field - a logically distinguishable area within a record in which data
is located.
File - a collection of related records treated as a unit.
Floating-Point Binary - a method for storing numeric data,
the
exponent occupies the first byte and the mantissa resides in the next
three or seven bytes. Used in FORTRAN IV representation.
Format - the arrangement of any record or file;
fields reside in a record.

the

order

which

Key, Key Field - the sorting key is the data element chosen from a
record to control the sort. For example, if a list of names is sorted
alphabetically, the keys would be each successive letter in the name
field.
Key, Major - the most important field in the total key. If you were
sorting a list by name, department and salary, name would be the major
key.
Key, Minor - the least significant field in the
example above, salary is the minor key.

total

key.

the

Line - In this document, a line generally refers to a line in the SORT
specifications form or a record on the specification file.
Packed Decimal - 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.
RMS - Record Management Services
Glossary-!

Record - the unit of information in a file;
treated as a unit

a group of related fields

System Device - the device on which the Executive resides.
Word - two bytes or sixteen bits of data
Work File - a collection of sorted records created
processing cycle and released after the sort is finished

during

the

Zone - the top three punches of a character as represented on a
card or the numeric value of the upper four bits of a byte

punch

Glossary-2

INDEX

/AL,
/BL,
/BU,
/CO,
/DE,
/FI,
/FO,
/IN,
/KE,
/PR,

/RE,
/SE,
/SI,

3-1
3-1
3-1
3-3
3-3
3-3
3-3
3-3
3-4
3-4
3-4
3-4
3-4

Block,
512-byte, 3-1
Blocking,
big, D-2
Blocks per bucket, 3-1
Blocksize switch, 3-1
Boolean expression, 3-13
Bucket,
blocks per, 3-1
Bucket size, 3-1
Bucketsize switch, 3-1
Buffer size ratio,
scratch file, D-3
Byte,
null, 1-10
padding, 5-9

$SORT,
$RUN, 2-1
RUN, 2-2

Access from a random file,
direct, 1-1
sequential, 1-1
Address routing sort, 1-1, 1-8,
1-9, 1-10, 3-4
Allocation,
file, 5-2
Allocation switch, 3-1
Allocation, figure,
SORT work area, D-3
Alternate collating sequence,
1-6' 3-6' 3-7
Alternate key, 1-2
ALTSEQ, 1-6, 3-7
ALTSEQ format and notes, 3-10
ALTSEQ records, 3-6
ANSI format, 1-3
Ascending order, 1-4
ASCII, 1-10
ASCII character set, table, A-2
ASCII data, 1-2
ASCII zone, 3-4
Assignments table,
LUN, 5-1
Available internal storage
space, 1-8

Batch mode,
running SORT in RSTS/E, 2-2
Binary data, 1-2, 1-9

Calling sequence,
SORT subroutine, 5-10
Card reader, 1-2
Cards, 1-3
Changing SORT defaults, 3-1
Changing sort order within
a sequence, 3-7
Character set, table,
ASCII, A-2
EBCDIC, A-3
Characters, equal digits,
printable, table, B-2
Characters, equal zones,
printable, table, B-3
Cluster size,
file, 3-5
Code,
device, 2-3
user identification, 2-3
Codes,
RMS error, 4-7
RMS error status, F-1
SORT error, 5-5
Codes, table,
device specification, 2-4
Collating sequence,
alternate, 1-6, 3-6, 3-7
normal, 1-6
Collation,
nonstandard, 3-5
Command decoder errors, 4-1
Command file,
indirect, 2-2
Command string, 1-1, 1-4, 2-2
errors, 1-8
format, 2-1, 2-2

lndex-1

INDEX (CONT . )

Conunand string (Cont.),
functions, 1-4
SORT, 2-1
switches, 2-2
Conunents, 3-8
Conditional keys, 3-17
Conditions,
error, 4-1
Configuration,
system, 1-2
Console, 1-3
Contiguous switch, 3-3
Control fields, 1-1
Control program, 1-7
errors, 4-6
Control-phase, 4-1
Controlling SORT, 1-1
Controlling SORT options, 3-1
Controls,
sorting process, 3-5
Conventions,
file naming, 5-1

Data,
ASCII, 1-2
binary, 1-2, 1-9
EBCDIC, 1-2
entry, size switch defaults,
3-5
field, 1-2
defining, 3-7
locating, 3-7
specifications, 3-7
handling, 3-4
irrelevant to SORT, 1-2
key, 1-2
output, 1-9
relevant, 1-2
restricted-format key, 3-7
Data base, 1-1
Decoder errors,
conunand, 4-1
Defaults,
changing SORT, 3-1
file, footnote, 3-2
file types, 2-3
sort process, 1-8
switch, table, 3-2
Defining a data field, 3-7
Defining a key, 3-7
Description,
SORT process, D-1
Description sequence,
field, 3-4
Device,
code, 2-3
file, input and output, 2-2

Device (Cont.),
peripheral, 1-2
specification codes, table,
2-4
switch, 3-3
types, table, 2-4
Different record formats, 1-7
Digits,
printable characters,
equal, B-2
Direct access from a random
file, 1-1
Directing SORT, 1-5
Discussion,
specification file example,
3-17
Disk, 1-2
scratch files on, 1-8
wasted space on, 5-2
Distribution process,
record, D-3

EBCDIC,
character set, table, A-3
data, 1-2
ENDS, 5-9
Entry,
conunon, of specification
lines, 3-7
field specification, 3-15
forced key specification,
3-17
header specification, 3-9
notes, 3-10
points, SORTS subroutines,
5-6
record type specification,
3-17
SORT specification, figure,
3-18
Error,
codes,
RMS, 4-7
RMS status, F-1
SORT, 5-5
conunand decoder, 4-1
conunand string, 1-8
control program, 4-6
I/O, 4-1, 4-5
merge, 4-6
messages, 4-1
other, 4-6, 4-8
presort, 4-6
RSX-llM file primitive, 4-7
RSX-llM attribute mismatch,
4-7
SORT~ 4-1

Index-2

INDEX (CONT • )

Error (Cont.),
specification file, 1-8, 4-1,
4-3
subroutine, 4-8
Example,
SORT, C-1
SORTR, 3-17
specification file discussion,
3-17
Expression,
Boolean, 3-13

Factor,
big blocking, D-2
Field,
control, 1-1
data, 1-2
defining a data, 3-7
description sequence, 3-4
first byte, key, 3-4
interpretation, 1-2
locating a data, 3-7
key, 1-1, 1-5, 1-10, 3-4
length, key, 3-4
location, key, 1-5
specification, 3-7, 3-14
data, 3-7
entries, 3-15
key, 3-7
notes, 3-16
stripping key, 3-14
table, 3-22
File,
allocation, 5-2
buffer size ratio, scratch,
D-3
cluster size, 3-5
defaults, footnote, 3-2
devices,
input and output, 2-2
errors,
specification, 1-8, 4-1, 4-3
example, discussion,
specification, 3-17
index, 1-1
indexed, 1-2
indirect command, 2-2
information,
RMS, 1-5
input, 1-1, 1-9, 2-2, 3-17,
5-1
input and output, 3-1
internal work, 1-5
merging scratch, 1-8
moving records within a, 3-7
name, 2-3
naming conventions, 5-1

File (Cont.),
needed, number of scratch, 1-8
on disk, scratch, 1-8
opening the input, 1-8
optional specification, 1-1
organization,
input, 3-1
RMS-valid, 1-2
output, 1-8, 1-10, 1-11, 2-1,
2-2' 3-7' 5-1
parameters, 1-5
input and output, 2-2
primitive errors,
RSX-llM, 4-7
random, direct access, 1-1
relative, 1-2
scratch, 5-1
sequential, 1-2
sequential access from a
random, 1-1
sorted, 2-1
specification, 1-5, 1-8, 2-2,
2-3, 3-5, 3-14, 3-17, 4-3,
5-1
structure, scratch, 5-3
switch, 3-3
type, 2-3
types, default, 2-3
uniform format input, 1-5
usage, scratch, 5-3
First byte,
key field, 3-4
FIXED format, 3-3
FIXED records, 1-8
Forced key specification entries,
3-17
Forced keys, 1-6
Form,
figure, 3-8
SORT specification, 3-7
Format,
ALTSEQ, and notes, 3-10
ANSI, 1-3
command string, 2-1, 2-2
FIXED, 3-3
output, 1-7
packed decimal, 3-4
record, 1-5
different, 1-7
STREAM, 3-3
switch, 3-3
uniform input file, 1-5
UNKNOWN, 3- 3
VARIABLE, 3-3
variable record, 3-7
variation, input, 1-7
FORTRAN IV number, 3-4
Functions,
command string, 1-4

Index-3

INDEX (CONT.)

General data, irrelevant to
SORT, 1-2
Handling,
data, 3-4
Header specification, 3-6, 3-17
entries, 3-9
notes., 3-10
table, 3-19
High-speed sorting, 1-8

I/O errors, 4-1, 4-5
IAS, 2-1
Identification by type,
record, 1-7
Identification code,
user, 2-3
Identifier,
record, 1-2
Index,
file, 1-1
keys, 1-10
record, 1-1, 1-10, 1-11
sort, 1-1, 1-8, 1-10, 1-11, 3-4
Indexed file, 1-2
Indexed sequential switch, 3-3
Indices,
record, 1-11
Indirect command file, 2-2
Information,
key, 1-1
prologue, 1-2
RMS file, 1-5
storage of sorted, 1-1
Initial sorting process, 1-8
Input and output files, 3-1
devices, 2-2
parameters, 2-2
Input file, 1-1, 1-9, 2-2, 3-17,
5-1
opening the, 1-8
organization, 3-1
uniform format, 1-5
Input format variation, 1-7
Input phase, 1-9
Installation,
SORT, E-1
Interactive mode, 2-1
Internal operations,
SORT, 5-1, D-1
Internal work files, 1-5
Interpretation,
field, 1-2
Invoking SORT, 2-1
Irrelevant to SORT,
general data, 1-2

Key,
alternate, 1-2
area size, D-2
conditional, 3-17
data, 1-2
restricted-format, 3-7
defining, 3-7
fields, 1-1, 1-5, 1-10, 3-4
first byte, 3-4
length, 3-4
location, 1-5
specifications, 3-7
stripping, 3-14
forced, 1-6
specification entries, 3-17
in original form,
record, 1-11
index, 1-10
information, 1-1
locating a, 3-7
primary, 1-2
setup, 5-9
significance,
sort, 1-4
switch, 3-4

Labels,
magtape, 1-9
Length,
key field, 3-4
Line number, 3-7
Lines,
common entries of specif ication, 3-7
Locating a data field, 3-7
Locating a key, 3-7
Location,
key field, 1-5
Logical Unit Number, 5-1
LON assignments table, 5-1

Magtape, 1-2, 3-1
Magtape labels, 1-9
Management Services,
Record, 1-1, 1-8
Maximum record size switch, 1-5
MERGE, 5-8
Merge errors, 4-6
Merge phase, 1-8
Merging scratch files, 1-8
Messages,
SORT error, 4-1
Method,
sorting, 1-1

Index-4

INDEX (CONT • )

Mismatch errors,
RSX-llM, 4-7
Mode,
interactive, 2-1
Moving records within a file,
3-7

Output file, 1-8, 1-10, 1-11,
2-1, 2-2, 3-7
devices, 2-2
parameters, 2-2, 3-1, 5-1
Output format variation, 1-7
Output record, 1-11, 3-14

Name,
switch, 2-2
Naming conventions,
file, 5-1
Nonstandard collation, 3-5
Normal collating sequence, 1-6
Notes,
ALTSEQ format and, 3-10
field specification, 3-16
record type specifications,
3-13
Null byte, 1-10
Null records, 3-17
Number,
FORTRAN IV, 3-4
line, 3-7
logical unit, 5-1
page, 3-7
project-programmer, 2-3
Number of scratch files needed,
1-8

Package,
SORTS subroutine, 1-7, 5-5
Packed decimal format, 3-4
Padding byte, 5-9
Page number, 3-7
Paper tape, 1-2
Parameters,
file, 1-5
input and output file, 2-2
SORT, 3-4
SORT performance, D-4
SORT performance, table, D-5
Parts of SORT, 1-7
Performance parameters,
SORT, D-4
table, D-5
Peripheral device, 1-2
Phase,
input, 1-9
merge, 1-8
SORT, 1-7
Pointer,
record, 1-2
relative record, 1-10
Points,
SORTS subroutine entry, 5-6
Preallocation problems, 5-2
Presort errors, 4-6
Presort operation, 1-8
Primary key, 1-2
Printable characters, equal
digits, table, B-2
Printable characters, equal
zones, table, B-3
Printer, 1-3
Problems,
preallocation, 5-2
Process,
controls, sorting, 3-5
default sort, 1-8
description, SORT, D-1
initial sorting, 1-8
options, 1-5
table, 1-9
record distribution, D-3
SORT, 1-8
sorting, 1-5, 1-9
switch, 3-4

Opening the input file, 1-8
Operations,
SORT, 3-8
SORT internal, 5-1, D-1
Optimization,
sort, 5-2
Optional specification file, 1-1
Options,
controlling SORT, 3-1
sort processing, 1-8
sorting process, 1-5
sorting process, table, 1-9
Order, 1-4
ascending, 1-4
ascending sort, 3-4
descending sort, 3-4
general sort, 3-4
Organization,
input file, 3-1
SORT, 5-4
Organization, RMS-valid,
file, 1-2
Other errors, 4-6, 4-8
Output data, 1-9

Index-5

INDEX (CONT. )

Processing,
options, sort, 1-8
sort, 1-10
Program, control, 1-7
errors, 4-6
Project-programmer number, 2-3
Prologue information, 1-2
Prompt,
SORT, 2-2
system, 2-1

Random file,
direct access from a, 1-1
sequential access from a, 1-1
Ratio,
scratch file buffer size, D-3
Reel time, 2-1
Record Management Services, 1-1,
1-8
Records,
ALTSEQ, 3-6
distribution process, D-3
FIXED, 1-8
format,
different, 1-7
switch, 1-5
variable, 3-7
identification by type, 1-7
identifier, 1-2
index, 1-1, 1-10, 1-11
keys in original form, 1-11
moving, within a file, 3-7
null, 3-17
output, 1-11, 3-14
pointer, 1-2
selection, 1-5, 3-7
size switch, maximum,
1-5
sort, 1-1, 1-8, 3-4
sorted, 2-1
STREAM ASCII, 1-8
strings of sorted, 1-8
type specifications, 3-7
entries, 3-11
notes, 3-13
table, 3-20
VARIABLE length, 1-8
within a file, moving, 3-7
Relative file, 1-2
Relative record pointers, 1-10
Relative switch, 3-4
RELES, 5-7
Relevant data, 1-2
Restricted-format key data, 3-7
Retrieval window size, 3-5
RETRN, 5-8

RMS, 1-1, 1-8, 3-1, D-2
error codes, 4-7
error status codes, F-1
file information, 1-5
RMS-valid,
file organization, 1-2
RSORT, 5-4, 5-6, 5-9
RSTS/E, 2-1, 2-2
version number, omitted for,
2-3
RSX-llM, 2.-1
file primitive errors, 4-7
mismatch errors, 4-7
$RUN $SORT, 2-2
RUN [1,2] SORT, 2-1
Running SORT, 2-1
in RSTS/E batch mode, 2-2

Scratch files, 5-1
buffer size ratio, D-3
merging, 1-8
needed, number of, 1-8
on disk, 1-8
structure, 5-3
usage, 5-3
Selection,
record, 1-5, 3-7
Sequence, 1-11
alternate collating, 1-6, 3-6,
3-7
changing sort order within a,
3-7
field description, 3-4
normal collating, 1-6
SORT subroutine calling, 5-10
Sequential,
access from a random file,
1-1
file, 1-2
switch, 3-4
Significance,
sort key, 1-4
Size,
bucket, 3-1
file cluster, 3-5
key area, D-2
retrieval window, 3-5
Size ratio,
scratch file buffer, D-3
Size switch, 3-4
maximum record, 1-5
Size switch defaults, data
entry, 3-5
$SORT,
RUN I 2-1
$RUN I 2-2

Index-6

INDEX (CONT . )

SORT, 1-1, 1-4, 1-8, 2-1, 3-17,
5-1
address routing, 1-1, 1-8 to
1-10 I 3-4
command string, 2-1
controlling, 1-1
defaults, changing, 3-1
directing, 1-5
error codes, 5-5
error messages, 4-1
errors, 4-1
examples, c-1
general data, irrelevant to,
1-2
Index, 1-1
index, 1-8, 1-10, 1-11, 3-4
installation, E-1
internal operations, 5-1, D-1
invoking, 2-1
key significance, 1-4
operations, 3-8
optimization, 5-2
options, controlling, 3-1
order,
ascending, 3-4
descending, 3-4
general, 3-4
within a sequence, changing,
3-7
organization, 5-4
parameters, 3-4
performance parameters, D-4
parameters, performance,
table, D-5
parts of, 1-7
phases, 1-7
process, 1-8
default, 1-8
description, D-1
processes, 1-8
processing, 1-10
options, 1-8
prompt, 2-2
Record, 1-1
record, 1-8, 3-4
running, 2-1
in RSTS/E batch mode, 2-2
specification,
entries, figure, 3-18
form, 3-7
figure, 3-8
subroutine calling sequence,
5-10
switches, 3-1
Tag, 1-1, 1-8, 1-9, 3-4
work area, D-1
allocation, figure, D-3
SORTA, 1-1, 1-9, 1-10

SORTC I 5-4
Sorted records, 2-1
Sorted file, 2-1
Sorted information,
storage of, 1-1
Sorted records,
strings of, 1-8
SORT!, 1-1, 1-10, 1-11, 3-17
Sorting,
high-speed, 1-8
method, 1-1
process, 1-5, 1-9
controls, 3-5
initial, 1-8
options, 1-5
table, 1-9
SORTM, 5-4
SORTP, 5-4
SORTR, 1-1, 1-9, 3-14
example, 3-17
SORTS I 5-4
subroutine entry points, 5-6
subroutine package, 1-7, 5-5
SORTT, 1-1, 1-9, 3-14
Space,
available internal storage,
1-8
Space on disk,
wasted, 5-2
Specification,
device codes, table, 2-4
entries,
field, 3-15
notes, 3-16
forced key, 3-17
header, 3-9
notes, header, 3-10
record type, 3-11
SORT, figure, 3-18
field, 3-7, 3-14
file, 1-5, 1-8, 2-2, 2-3,
3-5, 3-14, 3-17, 4-3, 5-1
errors, 1-8, 4-1, 4-3
example, discussion, 3-17
optional, 1-1
switch, 1-5
form, SORT, 3-7
figure, 3-8
header, 3-6, 3-17
key field, 3-7
lines, common entries of, 3-7
record type, 3-7
notes, 3-13
table,
field, 3-22
header, 3-19
record type, 3-20
type, 3-7

Index-7

INDEX (CONT.)

Type,
file, 2-3
record identification by, 1-7
specification, 3-7
Type specifications,
record, 3-7
entries, 3-11
notes, 3-13
table, 3-20
Types,
default file, 2-3
Types, table,
device, 2-4

SRT I 2-1
Status codes,
RMS error, F-1
Storage of sorted information,
1-1
Storage space,
available internal, 1-8
STREAM ASCII records, 1-8
STREAM format, 3-3
String,
command, 1-1, 1-4, 2-2
of sorted records, 1-8
SORT command, 2-1
Stripping key fields, 3-14
Structure,
scratch file, 5-3
Subroutine calling sequence,
SORT I 5-10
Subroutine entry points,
SORTS I 5-6
Subroutine errors, 4-8
Subroutine package,
SORTS, 1-7, 5-5
Switch,
allocation, 3-1
blocksize, 3-1
bucketsize, 3-1
command string, 2-2
contiguous, 3-3
defaults, table, 3-2
device, 3-3
files, 3-3
format, 3-3
indexed sequential, 3~3
key I 3-4
maximum record size, 1-5
name, 2-2
process, 3-4
record format, 1-5
relative, 3-4
sequential, 3-4
size, 3-4
SORT I 3-1
specification, 1-5
System configuration, 1-2
System prompt, 2-1

UICI 2-3
Uniform format input file, 1-5
Unit Number,
Logical, 5-1
UNKNOWN file format, 3-3
Usage,
scratch file, 5-3
User identification code, 2-3

Valid terminators, 4-2
Values,
switch, 1-8, 3-1
VARIABLE format, 3-3, 3-7
VARIABLE length records, 1-8
Variation,
input format, 1-7
output format, 1-7
Version number, omitted for
RSTS/E, 2-3

Table, see also Table of Contents
EBCDIC character set, A-3
LUN assignments, 5-1
Tag sort, 1-1, 1-8, 1-9, 3-4
Terminal, 1-2
Terminators,
valid, 4-2
Time,
elapsed, 2-1
real, 2-1
wall clock, 2-1

Wall clock time, 2-1
Wasted space on disk, 5-2
Window size,
retrieval, 3-5
Work area,
SORT, D-1
Work area allocation,
SORT, figure, D-3
Work files,
internal, 1-5
WORKHI, D-2
WORKLO, D-2

Zone,
ASCII I 3-4
Zones,
printable characters,
equal, table, B-3

Index-8