This chapter includes the following information about using the SORT and MERGE statements to sort and merge records for sequential, line sequential (Alpha only), relative, and indexed files:

• Sorting data with the SORT statement ( Section 9.1)
• Merging data with the MERGE statement ( Section 9.2)
• Sample programs using the SORT and MERGE statements ( Section 9.3)

The SORT statement provides a wide range of sorting capabilities and options. To establish a SORT routine, you do the following:

1. Declare the sort file with an Environment Division SELECT statement.
2. Use a Data Division Sort Description (SD) entry to define the sort file's characteristics.
3. Use a Procedure Division SORT statement.

The following program segments demonstrate SORT program coding:

SELECT Statement (Environment Division)

SELECT SORT-FILE ASSIGN TO "SRTFIL"

An SD File Description Entry (Data Division)

SD SORT-FILE. 01 SORT-RECORD. 05 SORT-KEY1 PIC X(5). 05 SOME-DATA PIC X(25). 05 SORT-KEY2 PIC XX.

SORT Statement (Procedure Division)

SORT SORT-FILE ASCENDING KEY S-NAME USING NAME-FILE GIVING NEW-FILE.

The SORT statement names a sort file, sort keys, an input file, and an output file. An explanation of sort keys follows.

Sorting Concepts

Records are sorted based on the data values in the sort keys. Sort keys identify the location of a record or the ordering of data. The following example depicts unsorted employee name and address records used for creating mailing labels:

Smith,	Joe	234 Ash St.	New Boston	NH	04356
Jones,	Bill	12 Birch St.	Gardner	MA	01430
Baker,	Tom	78 Oak St.	Ayer	MA	01510
Thomas,	Pete	555 Maple St.	Maynard	MA	01234
Morris,	Dick	21 Harris St.	Acton	ME	05670

If you sort the addresses in the previous example in ascending order using the zip code as the sort key, the mailing labels are printed in the order shown in the following example:

					SORT KEY
Thomas,	Pete	555 Maple St.	Maynard	MA	01234
Jones,	Bill	12 Birch St.	Gardner	MA	01430
Baker,	Tom	78 Oak St.	Ayer	MA	01510
Smith,	Joe	234 Ash St.	New Boston	NH	04356
Morris,	Dick	21 Harris St.	Acton	ME	05670

Also, records can be sorted on more that one key at a time. If you need an alphabetical listing of all employees within each state, you can sort on the state code first (major sort key) and employee name second (minor sort key).

For example, if you sort the file in ascending order by state and last name, the employee names and addresses appear in the order shown in the following example:

SORT KEY (minor)				SORT KEY (major)
Baker,	Tom	78 Oak St.	Ayer	MA	01510
Jones,	Bill	12 Birch St.	Gardner	MA	01430
Thomas,	Pete	555 Maple St.	Maynard	MA	01234
Morris,	Dick	21 Harris St.	Acton	ME	05670
Smith,	Joe	234 Ash St.	New Boston	NH	04356

9.1.1 File Organization Considerations for Sorting

You can sort any file regardless of its organization; furthermore, the organization of the output file can differ from that of the input file. For example, a sort can have a sequential input file and a relative output file. In this case, the relative key for the first record returned from the sort is 1; the second record's relative key is 2; and so forth. However, if an indexed file is described as output in the GIVING or OUTPUT PROCEDURE phrases, the first sort key associated with the ASCENDING phrase must specify the same character positions specified by the RECORD KEY phrase for that file.

Sections 9.1.2, 9.1.3, and 9.1.4 describe the ASCENDING and DESCENDING KEY phrases, the USING and GIVING phrases, and the INPUT PROCEDURE and OUTPUT PROCEDURE phrases for sorting.

9.1.2 Specifying Sort Parameters with the ASCENDING and DESCENDING KEY Phrases

Use the Data Division ASCENDING and DESCENDING KEY phrases to specify your sort parameters. The order of data names determines the sort hierarchy; that is, the major sort key is the first data name entered, while the minor sort key is the last data name entered.

In the following example, the hierarchy of the sort is SORT-KEY-1, SORT-KEY-2, SORT-KEY-3.

SORT SORT-FILE ASCENDING KEY SORT-KEY-1 SORT-KEY-2 DESCENDING KEY SORT-KEY-3

9.1.3 Resequencing Files with the USING and GIVING Phrases

If you only need to resequence a file, use the USING and GIVING phrases of the SORT statement. The USING phrase opens the input file, then reads and releases its records to the sort. The GIVING phrase opens and writes sorted records to the output file.

Note that you cannot manipulate data with either the USING or the GIVING phrases.

Consider this SORT statement:

SORT SORT-FILE ON ASCENDING KEY SORT-KEY-1 USING INPUT-FILE GIVING OUTPUT-FILE.

It does the following:

1. Opens INPUT-FILE
2. Reads all records in INPUT-FILE and releases them to the sort
3. Sorts the records in ascending sequence using the data in SORT-KEY-1
4. Opens the output file and writes the sorted records to OUTPUT-FILE
5. Closes all files used in the SORT statement

You can manipulate data before and after sorting by using the INPUT PROCEDURE and OUTPUT PROCEDURE phrases, and sort only some of the information in a file. For example, these phrases allow you to use only those input records and/or input data fields you need.

The INPUT PROCEDURE phrase replaces the USING phrase when you want to manipulate data entering the sort. The SORT statement transfers control to the sections or paragraphs named in the INPUT PROCEDURE phrase. You then use COBOL statements to open and read files, and manipulate the data. You use the RELEASE statement to transfer records to the sort. After the last statement of the input procedure executes, control is given to the sort, and the records are subsequently sorted.

After the records are sorted, the SORT statement transfers control to the sections or paragraphs named in the OUTPUT PROCEDURE phrase. This phrase replaces the GIVING phrase when you want to manipulate data in the sort. You can use COBOL statements to open files and manipulate data. You use the RETURN statement to transfer records from the sort. For example, you can use the RETURN statement to retrieve the sorted records for printing a report.

Example 9-1 shows a sample sort using the INPUT and OUTPUT procedures.

Example 9-1 INPUT and OUTPUT PROCEDURE Phrases

IDENTIFICATION DIVISION. PROGRAM-ID. EX0901. ENVIRONMENT DIVISION. INPUT-OUTPUT SECTION. FILE-CONTROL. SELECT INPUT-FILE ASSIGN TO "input.dat". SELECT OUTPUT-FILE ASSIGN TO "output.dat". SELECT SORT-FILE ASSIGN TO "sort.dat". DATA DIVISION. FILE SECTION. FD INPUT-FILE. 01 INPUT-RECORD PIC X(100). FD OUTPUT-FILE. 01 OUTPUT-RECORD PIC X(100). SD SORT-FILE. 01 SORT-RECORD PIC X(100). 01 SORT-KEY-1 PIC XXX. 01 SORT-KEY-2 PIC XXX. WORKING-STORAGE SECTION. PROCEDURE DIVISION. 000-SORT SECTION. 010-DO-THE-SORT. SORT SORT-FILE ON ASCENDING KEY SORT-KEY-1 ON DESCENDING KEY SORT-KEY-2 INPUT PROCEDURE IS 050-RETRIEVE-INPUT THRU 100-DONE-INPUT OUTPUT PROCEDURE IS 200-WRITE-OUTPUT THRU 230-DONE-OUTPUT. DISPLAY "END OF SORT". STOP RUN. 050-RETRIEVE-INPUT SECTION. 060-OPEN-INPUT. OPEN INPUT INPUT-FILE. 070-READ-INPUT. READ INPUT-FILE AT END CLOSE INPUT-FILE GO TO 100-DONE-INPUT. MOVE INPUT-RECORD TO SORT-RECORD. *********************************************************** * You can add, change, or delete records before sorting * * using COBOL data manipulation * * techniques. * *********************************************************** RELEASE SORT-RECORD. GO TO 070-READ-INPUT. 100-DONE-INPUT SECTION. 110-EXIT-INPUT. EXIT. 200-WRITE-OUTPUT SECTION. 210-OPEN-OUTPUT. OPEN OUTPUT OUTPUT-FILE. 220-GET-SORTED-RECORDS. RETURN SORT-FILE AT END CLOSE OUTPUT-FILE GO TO 230-DONE-OUTPUT. MOVE SORT-RECORD TO OUTPUT-RECORD. *********************************************************** * You can add, change, or delete sorted records * * using COBOL data manipulation * * techniques. * *********************************************************** WRITE OUTPUT-RECORD. GO TO 220-GET-SORTED-RECORDS. 230-DONE-OUTPUT SECTION. 240-EXIT-OUTPUT. EXIT.

You can combine the INPUT PROCEDURE with the GIVING phrases, or the USING with the OUTPUT PROCEDURE phrases. In Example 9-2, the USING phrase replaces the INPUT PROCEDURE phrase used in Example 9-1.

Example 9-2 USING Phrase Replaces INPUT PROCEDURE Phrase

. . . PROCEDURE DIVISION. 000-SORT SECTION. 010-DO-THE-SORT. SORT SORT-FILE ON ASCENDING KEY SORT-KEY-1 ON DESCENDING KEY SORT-KEY-2 USING INPUT-FILE OUTPUT PROCEDURE IS 200-WRITE-OUTPUT THRU 230-DONE-OUTPUT. DISPLAY "END OF SORT". STOP RUN. 200-WRITE-OUTPUT SECTION. 210-OPEN-OUTPUT. OPEN OUTPUT OUTPUT-FILE. 220-GET-SORTED-RECORDS. RETURN SORT-FILE AT END CLOSE OUTPUT-FILE GO TO 230-DONE-OUTPUT. MOVE SORT-RECORD TO OUTPUT-RECORD. WRITE OUTPUT-RECORD. GO TO 220-GET-SORTED-RECORDS. 230-DONE-OUTPUT SECTION. 240-EXIT-OUTPUT. EXIT.

9.1.5 Maintaining the Input Order of Records Using the WITH DUPLICATES IN ORDER Phrase

The sort orders data in the sequence specified in the ASCENDING KEY and DESCENDING KEY phrases. However, records with duplicate sort keys may not be written to the output file in the same sequence as they were read into it. The WITH DUPLICATES IN ORDER phrase ensures that any records with duplicate sort keys are in the same order in the output file as in the input file.

The following list shows the potential difference between sorting with the WITH DUPLICATES IN ORDER phrase and sorting without it:

Input File	Sorted Without Duplicates in Order	Sorted With Duplicates in Order
Record	Record	Record
Name Data	Name Data	Name Data
JONES ABCD	DAVIS LMNO	DAVIS LMNO
DAVIS LMNO	JONES EFGH	JONES ABCD
WHITE STUV	JONES ABCD	JONES EFGH
JONES EFGH	SMITH 1234	SMITH 1234
SMITH 1234	WHITE STUV	WHITE STUV
WHITE WXYZ	WHITE WXYZ	WHITE WXYZ

If you omit the WITH DUPLICATES IN ORDER phrase, you cannot predict the order of records with duplicate sort keys. For example, the JONES records might not be in the same sequence as they were in the input file, but the WHITE records might be in the same order as in the input file.

In contrast, the WITH DUPLICATES IN ORDER phrase guarantees that records with duplicate sort keys remain in the same sequence as they were in the input file.

9.1.6 Specifying Non-ASCII Collating Sequences with the COLLATING SEQUENCE IS Alphabet-Name Phrase

This phrase lets you specify a collating sequence other than the ASCII default. You define collating sequences in the Environment Division SPECIAL-NAMES paragraph. A sequence specified in the COLLATING SEQUENCE IS phrase of the SORT statement overrides a sequence specified in the Environment Division PROGRAM COLLATING SEQUENCE IS phrase.

Example 9-3 shows the alphabet name NEWSEQUENCE overriding the EBCDIC-CODE collating sequence.

Example 9-3 Overriding the COLLATING SEQUENCE IS Phrase

ENVIRONMENT DIVISION. OBJECT-COMPUTER. FOO PROGRAM COLLATING SEQUENCE IS EBCDIC-CODE. SPECIAL-NAMES. ALPHABET NEWSEQUENCE IS "ZYXWVUTSRQPONMLKJIHGFEDCBA" ALPHABET EBCDIC-CODE IS EBCDIC. . . . PROCEDURE DIVISION. 000-DO-THE-SORT. SORT SORT-FILE ON ASCENDING KEY SORT-KEY-1 SORT-KEY-2 COLLATING SEQUENCE IS NEWSEQUENCE USING INPUT-FILE GIVING OUTPUT-FILE.

9.1.7 Multiple Sorting

A program can contain multiple sort files, multiple SORT statements, or both multiple sort files and multiple SORT statements. Example 9-4 uses two sort files to produce two reports with different sort sequences.

Example 9-4 Using Two Sort Files

. . . DATA DIVISION. FILE SECTION. SD SORT-FILE1. 01 SORT-REC-1. 03 S1-KEY-1 PIC X(5). 03 FILLER PIC X(40). 03 S1-KEY-2 PIC X(5). 03 FILLER PIC X(50). SD SORT-FILE2. 01 SORT-REC-2. 01 SORT-REC-2. 03 FILLER PIC X(20). 03 S2-KEY-1 PIC X(10). 03 FILLER PIC X(10). 03 S2-KEY-2 PIC X(10). 03 FILLER PIC X(50). . . . PROCEDURE DIVISION. 000-SORT SECTION. 010-DO-FIRST-SORT. SORT SORT-FILE1 ON ASCENDING KEY S1-KEY-1 S1-KEY-2 WITH DUPLICATES IN ORDER USING INPUT-FILE OUTPUT PROCEDURE IS 050-CREATE-REPORT-1 THRU 300-DONE-REPORT-1. 020-DO-SECOND-REPORT. SORT SORT-FILE2 ON ASCENDING KEY S2-KEY-1 ON DESCENDING KEY S2-KEY-2 USING INPUT-FILE OUTPUT PROCEDURE IS 400-CREATE-REPORT-2 THRU 700-DONE-REPORT-2. 030-END-JOB. DISPLAY "PROGRAM ENDED". STOP RUN. 050-CREATE-REPORT-1 SECTION. ********************************************************** * * * * * Use the RETURN statement to read the sorted records. * * * * * ********************************************************** 300-DONE-REPORT-1 SECTION. 310-EXIT-REPORT-1. EXIT. 400-CREATE-REPORT-2 SECTION. ********************************************************** * * * * * Use the RETURN statement to read the sorted records. * * * * * ********************************************************** 700-DONE-REPORT-2 SECTION. 710-EXIT-REPORT. EXIT.

9.1.8 Sorting Variable-Length Records

If you specify the USING phrase and the input file contains variable-length records, the sort-file record must not be smaller than the smallest record, nor larger than the largest record, described in the input file.

If you specify the GIVING phrase and the output file contains variable-length records, the sort-file record must not be smaller than the smallest record, nor larger than the largest record, described in the output file.

9.1.9 Preventing I/O Aborts

All I/O errors detected during a sort can cause abnormal program termination. The Declarative USE AFTER STANDARD ERROR PROCEDURE, shown in Example 9-5, specifies error-handling procedures should I/O errors occur.

Example 9-5 The Declarative USE AFTER STANDARD ERROR PROCEDURE

PROCEDURE DIVISION. DECLARATIVES. SORT-FILE SECTION. USE AFTER STANDARD ERROR PROCEDURE ON INPUT-FILE. SORT-ERROR. DISPLAY "I-O TYPE ERROR WHILE SORTING". DISPLAY "INPUT-FILE STATUS IS " INPUT-STATUS. STOP RUN. END DECLARATIVES. 000-SORT SECTION. 010-DO-THE-SORT. SORT SORT-FILE ON DESCENDING KEY S-KEY-1 WITH DUPLICATES IN ORDER USING INPUT-FILE GIVING OUTPUT-FILE. DISPLAY "END OF SORT". STOP RUN.

The SORT statement can be used to order the elements in a table. This is especially useful for tables used with SEARCH ALL. The table elements are sorted based on the keys as specified in the OCCURS for the table unless you override them by specifying keys in the SORT statement. If no key is specified, the table elements are the SORT keys.

For the syntax and examples of table sorting, refer to the SORT statement description in the Procedure Division chapter of the Compaq COBOL Reference Manual. <>