Controlling the alphanumeric collating sequence with a locale

If you specify PROGRAM COLLATING SEQUENCE in the source program, the collating sequence is set at compile time and is used regardless of the locale at run time. If instead you set the collating sequence by using the COLLSEQ compiler option, the locale at run time takes precedence.

If the code page in effect is a single-byte ASCII code page, you can specify the following clauses in the SPECIAL-NAMES paragraph:

• ALPHABET clause
• SYMBOLIC CHARACTERS clause
• CLASS clause

If you specify these clauses when the source code page in effect includes DBCS characters, the clauses will be diagnosed and treated as comments. The rules of the COBOL user-defined alphabet-name and symbolic characters assume a character-by-character collating sequence, not a collating sequence that depends on a sequence of multiple characters.

If you specify the PROGRAM COLLATING SEQUENCE clause in the OBJECT-COMPUTER paragraph, the collating sequence that is associated with the alphabet-name is used to determine the truth value of alphanumeric comparisons. The PROGRAM COLLATING SEQUENCE clause also applies to sort and merge keys of USAGE DISPLAY unless you specify the COLLATING SEQUENCE phrase in the SORT or MERGE statement.

If you do not specify the COLLATING SEQUENCE phrase or the PROGRAM COLLATING SEQUENCE clause, the collating sequence in effect is NATIVE by default, and it is based on the active locale setting. This setting applies to SORT and MERGE statements and to the program collating sequence.

The collating sequence affects the processing of the following items:

• ALPHABET clause (for example, literal-1 THRU literal-2)
• SYMBOLIC CHARACTERS specifications
• VALUE range specifications for level-88 items, relation conditions, and SORT and MERGE statements