COBOL sort operations

COBOL sort operations let you put file records in order by one or more key fields. The SORT verb reads an input file, sorts the records (ascending or descending), and writes them to an output file. The MERGE verb combines two or more already-sorted files into one sorted file. Both use a special sort work file (SD) that the run time manages for you. For input/output procedures and RELEASE/RETURN, see Sorting operations.

The three files in a SORT

You define the sort work file with SD (Sort Description) in the File Section. Its record layout must match the records you are sorting. You never OPEN, READ, WRITE, or CLOSE the work file; the SORT statement does that internally.

Basic SORT syntax

The simplest form uses USING to name the input file and GIVING to name the output file. You specify one or more keys with ON ASCENDING KEY or ON DESCENDING KEY. The key fields must be part of the sort record.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
       FILE SECTION.
       FD  INPUT-FILE.
       01  INPUT-REC.
           05  IN-KEY    PIC X(10).
           05  IN-DATA   PIC X(70).
 
       SD  SORT-WORK.
       01  SORT-REC.
           05  SR-KEY    PIC X(10).
           05  SR-DATA   PIC X(70).
 
       FD  OUTPUT-FILE.
       01  OUTPUT-REC.
           05  OUT-KEY   PIC X(10).
           05  OUT-DATA  PIC X(70).
 
       PROCEDURE DIVISION.
           SORT SORT-WORK
               ON ASCENDING KEY SR-KEY
               USING INPUT-FILE
               GIVING OUTPUT-FILE
           END-SORT
           STOP RUN.

SORT-WORK is the sort work file (SD). SR-KEY is the sort key. The run time reads from INPUT-FILE, sorts by SR-KEY, and writes to OUTPUT-FILE. Input and output records must be the same length/structure as the sort record (or at least compatible for the compiler).

Multi-key SORT

You can sort by several keys. The first key is the primary key; later keys break ties. For example, sort by department (ascending), then by name (ascending), then by amount (descending).

1
2
3
4
5
6
7
       SORT SORT-WORK
           ON ASCENDING KEY SR-DEPT
           ON ASCENDING KEY SR-NAME
           ON DESCENDING KEY SR-AMT
           USING INPUT-FILE
           GIVING OUTPUT-FILE
       END-SORT.

MERGE: combining sorted files

MERGE is used when you have two or more files that are already sorted by the same key. It reads from all inputs, compares keys, and writes the next smallest (or largest) record to the output. The input files must be sorted; MERGE does not sort them, it only merges.

1
2
3
4
5
       MERGE SORT-WORK
           ON ASCENDING KEY SR-KEY
           USING INPUT-FILE-1, INPUT-FILE-2
           GIVING OUTPUT-FILE
       END-MERGE.

Step-by-step: simple SORT

• Define the input file (FD) and its record. Define the sort work file (SD) with the same (or compatible) record layout. Define the output file (FD) and its record.
• In the Procedure Division, write SORT sort-work-file ON ASCENDING (or DESCENDING) KEY key-name USING input-file GIVING output-file END-SORT.
• Do not OPEN or CLOSE the sort work file. You may need to OPEN/CLOSE input and output depending on your compiler; some handle it for USING/GIVING.

Explain like I'm five

SORT is like sorting cards: you have a pile (input), you use a temporary table (work file) to arrange them by number or name, then you put the sorted pile into a new box (output). MERGE is when you already have two sorted piles and you combine them into one sorted pile by always taking the smallest (or largest) card from the top of either pile.

Related concepts