COBOL COLLATING Sequence

The COLLATING SEQUENCE clause in COBOL defines the order in which characters are compared and sorted. It allows you to override default system ordering with custom alphabets, enabling business-specific sorting rules and international character handling.

Overview and Purpose

The COLLATING SEQUENCE clause provides powerful control over how characters are ordered in comparison and sorting operations. By default, COBOL uses the system's native collating sequence (ASCII on most systems, EBCDIC on mainframes), but business requirements often demand different ordering rules. For example, you might need to sort names with special characters in a specific way, or implement locale-specific alphabetical ordering.

Understanding collating sequences is crucial for data processing applications where sort order affects business logic, report generation, or user interface presentation. The ability to customize character ordering ensures that your applications can meet specific business requirements while maintaining data integrity and user expectations.

Basic Syntax and Definition

1
2
3
4
SPECIAL-NAMES.
    ALPHABET CUSTOM-ORDER IS
        "0" THRU "9"
        "A" THRU "Z".

This basic example defines a custom alphabet where digits (0-9) are ordered before letters (A-Z). The THRU keyword creates a range of characters in their natural sequence. This alphabet definition can then be referenced in COLLATING SEQUENCE clauses to control how data is sorted and compared. This particular ordering might be useful for product codes or identifiers where numeric portions should sort before alphabetic portions.

Sort Operation with Custom Collating

1
2
3
SORT SORT-FILE
    ON ASCENDING KEY SORT-KEY
    COLLATING SEQUENCE IS CUSTOM-ORDER.

This sort statement applies the custom collating sequence to determine record ordering. Instead of using the default system ordering, the sort operation will arrange records according to the CUSTOM-ORDER alphabet definition. This ensures that the sorted output meets specific business requirements, such as having all numeric entries appear before alphabetic entries, regardless of the underlying character encoding system.

International Character Ordering

1
2
3
4
ALPHABET INTERNATIONAL IS
    "A" "À" "Á" "Â" "Ã"
    "E" "È" "É" "Ê" "Ë"
    "N" "Ñ".

This example demonstrates how to define proper ordering for international characters with accents and diacritical marks. The alphabet groups base letters with their accented variants, ensuring that names like "García" and "Garcia" sort together rather than being separated by the entire alphabet. This type of collating sequence is essential for applications serving international markets or processing multilingual data.

Case-Insensitive Collating Sequence

1
2
3
ALPHABET CASE-INSENSITIVE IS
    "A" "a" "B" "b" "C" "c"
    "D" "d" "E" "e".

This alphabet definition creates case-insensitive ordering by pairing uppercase and lowercase letters together. When this collating sequence is used, "Apple", "apple", and "APPLE" will sort together rather than being separated by case differences. This approach is particularly useful for user-facing applications where case variations in data entry shouldn't affect the logical ordering of information.

Business-Specific Character Ordering

1
2
3
4
ALPHABET PRODUCT-CODES IS
    "0" THRU "9"
    "A" THRU "Z"
    "-" "_" ".".

This business-specific alphabet defines ordering for product codes that might contain digits, letters, and special separators. By explicitly defining the order of these characters, you ensure consistent sorting behavior regardless of the underlying system's default character ordering. This is particularly important when product codes follow specific formatting conventions that need to be preserved in sorted output.

File Processing with Custom Collating

1
2
3
4
5
6
SELECT CUSTOMER-FILE
    ASSIGN TO 'CUSTOMERS'
    ORGANIZATION IS INDEXED
    ACCESS MODE IS DYNAMIC
    RECORD KEY IS CUSTOMER-NAME
    COLLATING SEQUENCE IS INTERNATIONAL.

This file definition applies a custom collating sequence to an indexed file's record key. The INTERNATIONAL alphabet will control how customer names are ordered in the index, ensuring proper alphabetical sorting even when names contain accented characters or special symbols. This is crucial for maintaining logical data organization and efficient key-based retrieval operations in international business applications.

String Comparison with Collating Sequence

1
2
3
4
5
6
IF CUSTOMER-NAME-1 > CUSTOMER-NAME-2
    COLLATING SEQUENCE IS INTERNATIONAL
    PERFORM PROCESS-SECOND-CUSTOMER
ELSE
    PERFORM PROCESS-FIRST-CUSTOMER
END-IF.

This comparison operation uses a custom collating sequence to determine the relative ordering of two customer names. The INTERNATIONAL alphabet ensures that the comparison follows proper linguistic rules rather than simple ASCII or EBCDIC ordering. This is essential for implementing correct alphabetical sorting in user interfaces and reports where the order must make sense to human users rather than just following computer character encoding conventions.

Performance Monitoring for Collating Operations

1
2
3
4
01  WS-SORT-METRICS.
    05  WS-RECORDS-SORTED    PIC 9(8) VALUE ZERO.
    05  WS-COMPARISON-COUNT  PIC 9(10) VALUE ZERO.
    05  WS-SORT-START-TIME   PIC 9(8).

This performance monitoring structure tracks metrics related to sorting operations that use custom collating sequences. Custom collating can impact performance compared to default system ordering, so monitoring these metrics helps identify potential bottlenecks in large-scale data processing operations. The comparison count is particularly useful for understanding the computational overhead of custom character ordering rules.

Dynamic Collating Sequence Selection

1
2
3
4
5
6
7
8
EVALUATE WS-LOCALE-CODE
    WHEN 'EN-US'
        MOVE 'ENGLISH-ORDER' TO WS-COLLATING-NAME
    WHEN 'ES-ES'
        MOVE 'SPANISH-ORDER' TO WS-COLLATING-NAME
    WHEN 'FR-FR'
        MOVE 'FRENCH-ORDER' TO WS-COLLATING-NAME
END-EVALUATE.

This logic demonstrates how to select different collating sequences based on locale or user preferences. While the actual alphabet definitions must be compiled into the program, you can implement conditional logic to choose which predefined collating sequence to use for different operations. This approach enables applications to adapt their sorting behavior to different regional requirements or user preferences.

Error Handling for Collating Operations

1
2
3
4
5
6
7
8
9
10
SORT SORT-FILE
    ON ASCENDING KEY SORT-KEY
    COLLATING SEQUENCE IS CUSTOM-ORDER
    INPUT PROCEDURE IS SORT-INPUT-PROC
    OUTPUT PROCEDURE IS SORT-OUTPUT-PROC.
 
IF SORT-RETURN NOT = ZERO
    DISPLAY 'Sort operation failed with code: ' SORT-RETURN
    PERFORM ERROR-RECOVERY
END-IF.

This example shows proper error handling for sort operations that use custom collating sequences. Sort operations can fail for various reasons, including invalid collating sequence definitions or resource constraints. Checking the sort return code and implementing appropriate error recovery ensures that collating sequence problems don't cause silent data corruption or unexpected program termination.

Tutorial: Implementing Custom Sort Orders

1
2
3
4
5
SPECIAL-NAMES.
    ALPHABET BUSINESS-ORDER IS
        "0" THRU "9"
        "A" THRU "Z"
        SPACE.

1
2
3
4
5
SORT WORK-FILE
    ON ASCENDING KEY PRODUCT-CODE
    COLLATING SEQUENCE IS BUSINESS-ORDER
    USING INPUT-FILE
    GIVING OUTPUT-FILE.

1
2
3
4
5
6
OPEN INPUT OUTPUT-FILE
READ OUTPUT-FILE
PERFORM UNTIL END-OF-FILE
    DISPLAY PRODUCT-CODE
    READ OUTPUT-FILE
END-PERFORM.

Practical Exercises

1
2
3
4
5
6
7
8
ALPHABET INTERNATIONAL-NAMES IS
    "A" "À" "Á" "Â" "Ã" "Ä" "Å"
    "C" "Ç"
    "E" "È" "É" "Ê" "Ë"
    "I" "Ì" "Í" "Î" "Ï"
    "N" "Ñ"
    "O" "Ò" "Ó" "Ô" "Õ" "Ö"
    "U" "Ù" "Ú" "Û" "Ü".

1
2
3
4
ALPHABET INVENTORY-ORDER IS
    "A" "0" THRU "9" "-"
    "B" "0" THRU "9" "-"
    "C" "0" THRU "9" "-".

1
2
3
4
5
6
7
SEARCH CUSTOMER-TABLE
    AT END
        MOVE 'NOT-FOUND' TO SEARCH-RESULT
    WHEN CUSTOMER-NAME(INDEX-VAR) = SEARCH-NAME
        COLLATING SEQUENCE IS CASE-INSENSITIVE
        MOVE 'FOUND' TO SEARCH-RESULT
END-SEARCH.

Advanced Implementation Patterns

Locale-Specific Sorting

When developing applications for international markets, consider implementing multiple collating sequences for different locales. Each locale may have specific rules for character ordering that affect user expectations and data presentation requirements.

Performance Optimization

Custom collating sequences can impact sort performance, especially with large datasets. Consider the trade-offs between custom ordering requirements and processing speed. In some cases, post-processing sorted data might be more efficient than using complex collating sequences during the sort operation itself.

Maintenance and Documentation

Document your collating sequence decisions thoroughly, including the business rationale for specific character ordering choices. This documentation is crucial for maintenance and helps other developers understand why certain ordering rules were implemented.