cobol

Topics related to cobol:

Getting started with cobol

COBOL is the COmmon Business Oriented programming Language.

Even though it has become a pronounceable name, COBOL is still treated as an acronym by the standards committee, and COBOL is the preferred spelling by the ISO and INCITS standards bodies.

Standard Specification

The current specification is

ISO/IEC 1989:2014 Information technology – Programming languages, their environments and system software interfaces – Programming language COBOL

That document was published in May of 2014 and can be purchased from various branches of standard bodies, officially homed at

http://www.iso.org/iso/home/store/catalogue_tc/catalogue_detail.htm?csnumber=51416

Principal field of use

Business oriented. That usually means transaction processing. Banking, government agencies, and the insurance industry are major areas of COBOL application deployments. IBM mainframe systems usually have a COBOL compiler installed. There are upwards of 300 COBOL dialects in existence, with perhaps 10 or so versions taking the lion's share of deployments. Most of these compilers are proprietary systems, but free software COBOL is also available.

History

COBOL dates back to the late 1950s, with initial implementations published in 1960.

U.S. Navy Rear Admiral Grace Hopper is often associated with COBOL, and championed on behalf of the language during the early stages of development. She was not the only person involved in the design and development of COBOL, by any means, but is often referred to as the Mother of COBOL.

Due to early backing by governments and large corporations, COBOL has been in wide use for many decades. It remains a point of pride for some, and a thorn for others, who see it as outdated. The truth likely lies somewhere in between these extreme views. When applied to transaction processing, COBOL is at home. When applied to modern web screens and networking applications it may not feel as comfortable.

Structure

COBOL programs are written in four separate divisions.

IDENTIFICATION DIVISION
ENVIRONMENT DIVISION
DATA DIVISION
PROCEDURE DIVISION

Data Descriptions

Being designed to handle decimal data, COBOL allows for PICTURE based data descriptions, in grouped hierarchies.

01 record-group.
   05 balance        pic s9(8)v99.
   05 rate           pic 999v999.
   05 show-balance   pic $Z(7)9.99.

That defines balance as a signed eight digit value with two digits assumed after the decimal point. rate is three digits before and three digits after an assumed decimal point. show-balance is a numeric-edit field that will have a leading dollar sign, seven digits (zero suppressed) with at least one digit shown preceding two digits after a decimal point.

balance can be used in calculations, show-balance is only for display purposes and cannot be used in computational instructions.

Procedural statements

COBOL is a reserved keyword heavy language. MOVE, COMPUTE, MULTIPLY, PERFORM style long form words make up most of the standard specification. Over 300 keywords and 47 operational statements in the COBOL 2014 spec. Many compiler implementations add even more to the reserved word list.

GnuCOBOL installation with GNU/Linux

ACCEPT statement

The COBOL ACCEPT statement is used to retrieve data from the system.

ADD statement

Where rounded-phase is

ALLOCATE statement

Allocate working storage for a BASED item, or allocate a give size of heap storage.

ALTER statement

The much beloved ALTER statement. Changes the target of a GO TO paragraph.

No longer part of the COBOL standard, still supported by many compilers for reasons of backward compatibility. (The syntax diagram is dimmed to show that this is no longer standard COBOL).

CANCEL statement

The CANCEL statement ensures that a referenced program will be in an initial state the next time it is called, and to unload any resources for the module.

CALL statement

The COBOL CALL statement provides access to compiled library routines.

COMMIT statement

Flushes ALL current locks, synching file I/O buffers.

This is a non standard extension, available with some COBOL implementations that support ROLLBACK features.

COMPUTE statement

The COMPUTE statement allows for algebraic calculation expressions.

Rounded phrase is

CONTINUE statement

The CONTINUE statement causes the flow of control to continue at the next statement. Not quite a no-op, as it can influence control flow when inside compound statement sequences, in particular IF/THEN/ELSE.

A handy? example is during early development and building with and without debugging aids.

CALL "CBL_OC_DUMP" USING structure ON EXCEPTION CONTINUE END-CALL

That code, while expensive, will allow for formatted memory dumps when the module CBL_OC_DUMP is linked into the executable, but will harmlessly fail when it is not. *That trick is only applicable during early stages of development. The expense of a dynamic lookup failure is not something to leave in active code, and those lines should be removed from the source as soon as any initial concerns are satisfied in alpha testing. On first day coding, it can be a handy aid. By second day coding ON EXCEPTION CONTINUE occurrences should be wiped clean.

COPY directive

The COBOL version of the C #include preprocessor directive. Or, more historically accurate, COBOL came first, developed some 10 years earlier.

Due to some of the design decisions in COBOL (no arguments for PERFORM as the primary reason), many data structure access sequences need to break the DRY principle. Names of structure components need to be repeated in the ENVIRONMENT DIVISION, the DATA DIVISION and possibly many times in the PROCEDURE DIVISION. This is usually handled by adding copybooks. Record declarations and access code are tucked away in separate files and the COPY statement is the only repeated source. A change to the copybook keeps all uses of name spelling and data layout in synch, instead of requiring multiple edits to multiple files when a change occurs.

String

DELETE statement

The DELETE statement deletes records from mass storage. Some compilers allow the DELETE statement to be used with a FILE clause, to delete FD names (along with any associated indexing structures that may be required by the database management engine in use).

DIVIDE statement

The COBOL DIVIDE statement divides one numeric item into others setting data items to the quotient and, optionally, the remainder.

ROUNDED phrase:

Default is TRUNCATION if no rounded phrase specified. Default ROUNDED mode is NEAREST-TOWARD-ZERO (rounding down) unless other specified. So called Banker's rounding is NEAREST-EVEN.

DISPLAY statement

The DISPLAY statement causes data to be transferred to hardware or software of the operating environment. DISPLAY comes in two forms, UPON device or for display of SCREEN data. Environment variables can also be set with DISPLAY UPON in some implementations of COBOL, along with other extensions for data transfer of graphics or other device specific needs.

EVALUATE statement

The EVALUATE statement is a multiple branch, multiple join, conditional test and selection structure.

EXIT statement

The COBOL EXIT statement is a terminating flow control verb.

EXIT comes is a few flavours:

bare EXIT is a common end point for a series of procedures.
EXIT PARAGRAPH, EXIT SECTION provides a means of exiting a structured procedure without executing any of the subsequent statements.
EXIT FUNCTION, EXIT METHOD, EXIT PROGRAM marks the logical end of a module of code.
EXIT PERFORM breaks out of a inline perform loop.
EXIT PERFORM CYCLE causes an inline perform loop to begin the next iteration.

GENERATE statement

The COBOL GENERATE statement is an optional statement supported if the compiler includes the Report Writer feature.

FREE statement

The FREE statement frees allocated memory for one or more identifiers, either by POINTER or from a BASED working storage identifier. Use after FREE is illegal.

GO TO statement

The much beloved GO TO. COBOL includes named paragraphs and sections, along with other labels, and any of them can be the target of a GO statement.

GOBACK statement

The COBOL GOBACK statement is a return. Unlike EXIT PROGRAM, or STOP RUN, GOBACK always returns one level. If the current module is "main", GOBACK will return to the operating system. If the current module is a subprogram, GOBACK will return to the statement after a call.

IF statement

The conditional expression and selection statement. Use of explicit scope terminators is recommended. COBOL conditional expressions allow shortforms, where the current identifier (and conditional) is assumed through multiple condition tests, unless explicitly given.

IF A = 1 OR 2 ...

is equivalent to

IF A = 1 OR A = 2 ...

INITIALIZE statement

The INITIALIZE statement sets selected data to specified values.

Where category-name is:

INITIATE statement

The INITIATE statement initializes internal Report Writer control fields. Most of a report writer setup occurs in the DATA DIVISION with very brief PROCEDURE DIVISION statements. Once initialized, GENERATE does all the hard work of control break and paging of reports.

INSPECT statement

The INSPECT statement is a scan and replace verb in COBOL.

Where tallying-phrase is:

replacing-phrase is:

missing image

before-after-phrase is:

MERGE statement

The MERGE statement will merge one or more like formatted COBOL data files into a single output file. The programmer can assume control over theOUTPUT PROCEDURE, which uses the RELEASE statement, or use internal COBOL runtime mechanisms with the GIVING clause.

MOVE statement

MOVE is the workhorse of COBOL. Data is moved from literal or identifier to one or more identifiers. COBOL has a distinction between elementary and group MOVE. Elementary data is type converted from source to destination. Group data is moved as a byte array, without regard to field types with a structure. Numeric fields are moved from right to left, high order digit truncation with zero fill (normally). Alphanumeric character data is moved left to right, right end character truncation with space fill. There are quite a few rules on how MOVE goes about its business, with both BINARY and PICTURE DISPLAY data forms, and group hierarchies all accounted for.

MULTIPLY statement

The MULTIPLY statement multiplies numeric data setting the result to one or more identifiers of numeric type.

Where rounded-phrase is

OPEN statement

The COBOL OPEN statement initiates file processing. File resources in COBOL are defined in the ENVIRONMENT DIVISION, named in FD (File Descriptor) paragraphs. These fd names are used to access physical disk files and various options are specified in a SELECT clauses in the FILE-CONTROL paragraph of the INPUT-OUTPUT SECTION. A programmer is expected to test a FILE STATUS identifier for status and error codes.

Modes include INPUT, OUTPUT, I-O and EXTEND.

PERFORM statement

The PERFORM statement transfers control to one or more procedures and returns control implicitly when the sequence completes. PERFORM can also be used for inline loops withing the scope of the PERFORM.

The VARYING phrase allows for nesting with one or more AFTER clauses, and the conditional test can be BEFORE (default) or AFTER each loop.

The THRU clause of a procedural perform assumes sequential top down control flow from procedure-1 through the end of procedure-2. THRU is a contentious issue, and many programmers prefer PERFORM by SECTION rather than using THRU paragraphs. Some shops may mandate PERFORM THRU with an explicit exit point paragraph, others may ban the use of THRU finding it more difficult to debug.

Procedural perform:

Inline perform:

Where varying-phrase is:

READ statement

The READ statement is a staple of COBOL transaction processing programming. Reads data from external storage into working store. With or without locks or sharing, sequentially, by random access, or by key. Declarative clauses for AT END may also be specified, but some programmers prefer explicit FILE STATUS testing.

As each file resource may contain any type of record in any given slot, COBOL is a "read a file", "write a record" language, READ takes a filename (FD) and it is up to the programmer to place the record in an appropriate structure if heterogeneous data is saved in the file.

RELEASE statement

The RELEASE statement is used to give records to the COBOL SORT algorithm under programmer controlled conditions.

RETURN statement

The RETURN statement controls when data is sent to the internal COBOL sort algorithm writer, as part of an OUTPUT PROCEDURE. Post sort data can be transformed under programmer control before being returned and written to the output file by the sort algorithm.

REPLACE directive

The REPLACE directive is part of the COBOL standard preprocessor. Replacements are made before compilation begins.

REWRITE statement

The REWRITE statement logically replaces existing records on mass storage.

SET statement

The COBOL SET statement sets values, and operating environment data. It can be argued that SET was overused by the committee, as it has over a dozen documented syntax formats.

SEARCH statement

The COBOL SEARCH statement comes in two forms. Linear top to bottom SEARCH and a binary SEARCH ALL algorithm. Binary SEARCH ALL assumes a sorted table suitable for a binary search with no elements out of order.

SEARCH statement

Linear SEARCH

Binary SEARCH ALL

SORT statement

The COBOL SORT statement can be used to sort files and tables in working storage.

SORT file

SORT table

START statement

The START statement provides a way to position a read in a file for subsequent sequential retrieval (by key).

The key relational can include (but is not limited to):

KEY IS GREATER THAN
KEY IS >
KEY IS LESS THAN
KEY IS <
KEY IS EQUAL TO
KEY IS =
KEY IS NOT GREATER THAN
KEY IS NOT >
KEY IS NOT LESS THAN
KEY IS NOT <
KEY IS NOT EQUAL TO
KEY IS NOT =
KEY IS <>
KEY IS GREATER THAN OR EQUAL TO
KEY IS >=
KEY IS LESS THAN OR EQUAL TO
KEY IS <=

SUBTRACT statement

The SUBTRACT statement is used to subtract one, or the sum of two or more, numeric data items from one or more items, and set the values of one or more identifiers to the result.

rounded-phrase

STOP statement

The STOP statement terminates the current run unit.

A now deemed obsolete extension to STOP RUN is STOP literal, which will pause a program until a response from the console is given, at which point execution will resume. This could be handy for things like, "go get the big box of paper and load up the special printer".

STOP is a hard program end, GOBACK is a slightly nicer way of returning to the operating system or caller module, especially in subroutines that may have no business terminating a run.

TERMINATE statement

The TERMINATE statement is a COBOL Report Writer feature. Terminates the processing on the given report names.

STRING statement

The STRING statement concatenates the partial or complete contents of multiple fields into a single result.

SUPPRESS statement

The SUPPRESS statement inhibits the printing of a report group. COBOL Report Writer feature.

UNLOCK statement

The UNLOCK statement explicitly releases any record locks associated with a file connector.

Intrinsic Functions

COBOL 2014 lists the following standard Intrinsic Functions:

======================================== ==========
Intrinsic Function                       Parameters
======================================== ==========
FUNCTION ABS                             1
FUNCTION ACOS                            1
FUNCTION ANNUITY                         2
FUNCTION ASIN                            1
FUNCTION ATAN                            1
FUNCTION BOOLEAN-OF-INTEGER              2
FUNCTION BYTE-LENGTH                     1
FUNCTION CHAR                            1
FUNCTION CHAR-NATIONAL                   1
FUNCTION COMBINED-DATETIME               2
FUNCTION COS                             1
FUNCTION CURRENCY-SYMBOL                 0
FUNCTION CURRENT-DATE                    0
FUNCTION DATE-OF-INTEGER                 1
FUNCTION DATE-TO-YYYYMMDD                Variable
FUNCTION DAY-OF-INTEGER                  1
FUNCTION DAY-TO-YYYYDDD                  Variable
FUNCTION DISPLAY-OF                      Variable
FUNCTION E                               0
FUNCTION EXCEPTION-FILE                  0
FUNCTION EXCEPTION-FILE-N                0
FUNCTION EXCEPTION-LOCATION              0
FUNCTION EXCEPTION-LOCATION-N            0
FUNCTION EXCEPTION-STATEMENT             0
FUNCTION EXCEPTION-STATUS                0
FUNCTION EXP                             1
FUNCTION EXP10                           1
FUNCTION FACTORIAL                       1
FUNCTION FORMATTED-CURRENT-DATE          1
FUNCTION FORMATTED-DATE                  2
FUNCTION FORMATTED-DATETIME              Variable
FUNCTION FORMATTED-TIME                  Variable
FUNCTION FRACTION-PART                   1
FUNCTION HIGHEST-ALGEBRAIC               1
FUNCTION INTEGER                         1
FUNCTION INTEGER-OF-BOOLEAN              1
FUNCTION INTEGER-OF-DATE                 1
FUNCTION INTEGER-OF-DAY                  1
FUNCTION INTEGER-OF-FORMATTED-DATE       2
FUNCTION INTEGER-PART                    1
FUNCTION LENGTH                          1
FUNCTION LENGTH-AN                       1
FUNCTION LOCALE-COMPARE                  Variable
FUNCTION LOCALE-DATE                     2
FUNCTION LOCALE-TIME                     2
FUNCTION LOCALE-TIME-FROM-SECONDS        2
FUNCTION LOG                             1
FUNCTION LOG10                           1
FUNCTION LOWER-CASE                      1
FUNCTION LOWEST-ALGEBRAIC                1
FUNCTION MAX                             Variable
FUNCTION MEAN                            Variable
FUNCTION MEDIAN                          Variable
FUNCTION MIDRANGE                        Variable
FUNCTION MIN                             Variable
FUNCTION MOD                             2
FUNCTION MODULE-CALLER-ID                0
FUNCTION MODULE-DATE                     0
FUNCTION MODULE-FORMATTED-DATE           0
FUNCTION MODULE-ID                       0
FUNCTION MODULE-PATH                     0
FUNCTION MODULE-SOURCE                   0
FUNCTION MODULE-TIME                     0
FUNCTION MONETARY-DECIMAL-POINT          0
FUNCTION MONETARY-THOUSANDS-SEPARATOR    0
FUNCTION NATIONAL-OF                     Variable
FUNCTION NUMERIC-DECIMAL-POINT           0
FUNCTION NUMERIC-THOUSANDS-SEPARATOR     0
FUNCTION NUMVAL                          1
FUNCTION NUMVAL-C                        2
FUNCTION NUMVAL-F                        1
FUNCTION ORD                             1
FUNCTION ORD-MAX                         Variable
FUNCTION ORD-MIN                         Variable
FUNCTION PI                              0
FUNCTION PRESENT-VALUE                   Variable
FUNCTION RANDOM                          Variable
FUNCTION RANGE                           Variable
FUNCTION REM                             2
FUNCTION REVERSE                         1
FUNCTION SECONDS-FROM-FORMATTED-TIME     2
FUNCTION SECONDS-PAST-MIDNIGHT           0
FUNCTION SIGN                            1
FUNCTION SIN                             1
FUNCTION SQRT                            1
FUNCTION STANDARD-COMPARE                Variable
FUNCTION STANDARD-DEVIATION              Variable
FUNCTION STORED-CHAR-LENGTH              1
FUNCTION SUM                             Variable
FUNCTION TAN                             1
FUNCTION TEST-DATE-YYYYMMDD              1
FUNCTION TEST-DAY-YYYYDDD                1
FUNCTION TEST-FORMATTED-DATETIME         2
FUNCTION TEST-NUMVAL                     1
FUNCTION TEST-NUMVAL-C                   2
FUNCTION TEST-NUMVAL-F                   1
FUNCTION TRIM                            2
FUNCTION UPPER-CASE                      1
FUNCTION VARIANCE                        Variable
FUNCTION WHEN-COMPILED                   0
FUNCTION YEAR-TO-YYYY                    Variable
======================================== ==========

GnuCOBOL adds

======================================== ==========
FUNCTION CONCATENATE                     Variable
FUNCTION SUBSTITUTE                      Variable
FUNCTION SUBSTITUTE-CASE                 Variable
======================================== ==========

The keyword FUNCTION is required unless source (or compile time option) includes

ENVIRONMENT DIVISION.
CONFIGURATION SECTION.
REPOSITORY.
    FUNCTION ALL INTRINSIC.

Where ALL INTRINSIC can be a list of functions to be used without the FUNCTION prefix in PROCEDURE DIVISION statements.

The LENGTH function has a sorted history. Some compilers include a LENGTH reserved word. For GnuCOBOL, this reserved word is only recognized when used in the phrase LENGTH OF, the OF token is required to disambiguate the function from the older reserved word extension.

UNSTRING statement

The UNSTRING statement separates a sending field and places results in multiple receiving fields.

USE statement

The USE statement specifies procedures to be used

for error and exception handling in addition to those provided by other facilities
before printing of a designated report group
after detection of designated exception conditions

Obsolete usage includes specifying procedures to be used during DEBUGGING, and extensions include adding interstitial procedures for program start and end.

WRITE statement

The WRITE statement releases logical records to an output or input-output storage resource, and for logical positioning of lines within a logical page.

WRITE sequential

WRITE random

Data division

How does the computational work in cobol?