Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.matterai.so/llms.txt

Use this file to discover all available pages before exploring further.

COBOL, despite its longevity and importance in business applications, has several common anti-patterns that can lead to maintainability problems and bugs. Here are the most important anti-patterns to avoid when writing COBOL code.
* Anti-pattern: Excessive use of GO TO
PROCEDURE DIVISION.
   PERFORM INITIALIZE-ROUTINE.
   GO TO PROCESS-DATA.

INITIALIZE-ROUTINE.
   MOVE ZEROS TO TOTAL-AMOUNT.
   MOVE SPACES TO ERROR-MESSAGE.

PROCESS-DATA.
   READ INPUT-FILE
      AT END GO TO END-OF-FILE
   END-READ.
   
   IF AMOUNT-FIELD > 1000
      GO TO LARGE-AMOUNT-ROUTINE
   ELSE
      GO TO SMALL-AMOUNT-ROUTINE
   END-IF.

LARGE-AMOUNT-ROUTINE.
   ADD AMOUNT-FIELD TO LARGE-TOTAL.
   GO TO PROCESS-DATA.

SMALL-AMOUNT-ROUTINE.
   ADD AMOUNT-FIELD TO SMALL-TOTAL.
   GO TO PROCESS-DATA.

END-OF-FILE.
   DISPLAY "Processing complete".
   STOP RUN.

* Better approach: Structured programming
PROCEDURE DIVISION.
   PERFORM INITIALIZE-ROUTINE.
   PERFORM PROCESS-DATA UNTIL END-OF-FILE.
   DISPLAY "Processing complete".
   STOP RUN.

INITIALIZE-ROUTINE.
   MOVE ZEROS TO TOTAL-AMOUNT.
   MOVE SPACES TO ERROR-MESSAGE.

PROCESS-DATA.
   READ INPUT-FILE
      AT END SET END-OF-FILE TO TRUE
      NOT AT END PERFORM PROCESS-RECORD
   END-READ.

PROCESS-RECORD.
   IF AMOUNT-FIELD > 1000
      PERFORM LARGE-AMOUNT-ROUTINE
   ELSE
      PERFORM SMALL-AMOUNT-ROUTINE
   END-IF.

LARGE-AMOUNT-ROUTINE.
   ADD AMOUNT-FIELD TO LARGE-TOTAL.

SMALL-AMOUNT-ROUTINE.
   ADD AMOUNT-FIELD TO SMALL-TOTAL.
Avoid excessive use of GO TO statements, which create spaghetti code that is difficult to understand and maintain. Use structured programming constructs like PERFORM, IF-ELSE, and EVALUATE instead.
* Anti-pattern: Using magic numbers
IF CUSTOMER-TYPE = 1
   COMPUTE DISCOUNT = TOTAL-AMOUNT * 0.10
ELSE IF CUSTOMER-TYPE = 2
   COMPUTE DISCOUNT = TOTAL-AMOUNT * 0.15
ELSE IF CUSTOMER-TYPE = 3
   COMPUTE DISCOUNT = TOTAL-AMOUNT * 0.20
END-IF.

* Better approach: Use named constants
01 DISCOUNT-RATES.
   05 REGULAR-CUSTOMER-RATE     PIC 9V99 VALUE 0.10.
   05 PREFERRED-CUSTOMER-RATE   PIC 9V99 VALUE 0.15.
   05 PREMIUM-CUSTOMER-RATE     PIC 9V99 VALUE 0.20.

01 CUSTOMER-TYPES.
   05 REGULAR-CUSTOMER          PIC 9 VALUE 1.
   05 PREFERRED-CUSTOMER        PIC 9 VALUE 2.
   05 PREMIUM-CUSTOMER          PIC 9 VALUE 3.

* In PROCEDURE DIVISION
IF CUSTOMER-TYPE = REGULAR-CUSTOMER
   COMPUTE DISCOUNT = TOTAL-AMOUNT * REGULAR-CUSTOMER-RATE
ELSE IF CUSTOMER-TYPE = PREFERRED-CUSTOMER
   COMPUTE DISCOUNT = TOTAL-AMOUNT * PREFERRED-CUSTOMER-RATE
ELSE IF CUSTOMER-TYPE = PREMIUM-CUSTOMER
   COMPUTE DISCOUNT = TOTAL-AMOUNT * PREMIUM-CUSTOMER-RATE
END-IF.
Avoid using magic numbers (hardcoded numeric literals) in your code. Use named constants to make your code more readable and maintainable.
* Anti-pattern: Flat data structures
01 CUSTOMER-RECORD.
   05 CUSTOMER-ID           PIC X(10).
   05 CUSTOMER-NAME         PIC X(30).
   05 CUSTOMER-ADDRESS-1    PIC X(30).
   05 CUSTOMER-ADDRESS-2    PIC X(30).
   05 CUSTOMER-CITY         PIC X(20).
   05 CUSTOMER-STATE        PIC X(2).
   05 CUSTOMER-ZIP          PIC X(10).
   05 CUSTOMER-PHONE        PIC X(15).
   05 CUSTOMER-EMAIL        PIC X(50).

* Better approach: Use hierarchical data structures
01 CUSTOMER-RECORD.
   05 CUSTOMER-ID           PIC X(10).
   05 CUSTOMER-NAME         PIC X(30).
   05 CUSTOMER-ADDRESS.
      10 STREET-ADDRESS-1   PIC X(30).
      10 STREET-ADDRESS-2   PIC X(30).
      10 CITY               PIC X(20).
      10 STATE              PIC X(2).
      10 ZIP-CODE           PIC X(10).
   05 CUSTOMER-CONTACT.
      10 PHONE-NUMBER       PIC X(15).
      10 EMAIL-ADDRESS      PIC X(50).
Use hierarchical data structures to organize related data fields. This makes your code more readable and easier to maintain, especially when passing data between program modules.
* Anti-pattern: Hardcoded file names
SELECT CUSTOMER-FILE
   ASSIGN TO "C:\DATA\CUSTOMER.DAT"
   ORGANIZATION IS INDEXED
   ACCESS MODE IS DYNAMIC
   RECORD KEY IS CUSTOMER-ID.

* Better approach: Use configuration files or environment variables
SELECT CUSTOMER-FILE
   ASSIGN TO CUSTOMER-FILE-PATH
   ORGANIZATION IS INDEXED
   ACCESS MODE IS DYNAMIC
   RECORD KEY IS CUSTOMER-ID.

* In WORKING-STORAGE SECTION
01 CONFIGURATION-VARIABLES.
   05 CUSTOMER-FILE-PATH     PIC X(100).

* In PROCEDURE DIVISION
ACCEPT CUSTOMER-FILE-PATH FROM ENVIRONMENT "CUSTOMER_FILE_PATH".
IF CUSTOMER-FILE-PATH = SPACES
   MOVE "C:\DATA\CUSTOMER.DAT" TO CUSTOMER-FILE-PATH
END-IF.
Avoid hardcoding file paths and names in your programs. Use configuration files or environment variables instead, which makes your code more flexible and portable across different environments.
* Anti-pattern: Not validating input
ACCEPT CUSTOMER-ID.
READ CUSTOMER-FILE
   INVALID KEY DISPLAY "Customer not found"
END-READ.

* Better approach: Validate input data
ACCEPT CUSTOMER-ID.

* Validate input format
IF CUSTOMER-ID IS NOT NUMERIC
   DISPLAY "Error: Customer ID must be numeric"
   EXIT PARAGRAPH
END-IF.

IF CUSTOMER-ID = ZEROS
   DISPLAY "Error: Customer ID cannot be zero"
   EXIT PARAGRAPH
END-IF.

* Proceed with valid input
READ CUSTOMER-FILE
   INVALID KEY DISPLAY "Customer not found"
END-READ.
Always validate input data before processing it. This helps prevent runtime errors, data corruption, and potential security issues.
* Anti-pattern: Using PERFORM THRU
PROCEDURE DIVISION.
   PERFORM INITIALIZE-ROUTINE THRU INITIALIZE-EXIT.
   PERFORM PROCESS-DATA THRU PROCESS-EXIT.
   STOP RUN.

INITIALIZE-ROUTINE.
   MOVE ZEROS TO TOTAL-AMOUNT.
   MOVE SPACES TO ERROR-MESSAGE.
INITIALIZE-EXIT.
   EXIT.

PROCESS-DATA.
   READ INPUT-FILE
      AT END GO TO PROCESS-EXIT
   END-READ.
   ADD AMOUNT-FIELD TO TOTAL-AMOUNT.
   GO TO PROCESS-DATA.
PROCESS-EXIT.
   EXIT.

* Better approach: Use structured PERFORM statements
PROCEDURE DIVISION.
   PERFORM INITIALIZE-ROUTINE.
   PERFORM PROCESS-DATA UNTIL END-OF-FILE.
   STOP RUN.

INITIALIZE-ROUTINE.
   MOVE ZEROS TO TOTAL-AMOUNT.
   MOVE SPACES TO ERROR-MESSAGE.

PROCESS-DATA.
   READ INPUT-FILE
      AT END SET END-OF-FILE TO TRUE
      NOT AT END ADD AMOUNT-FIELD TO TOTAL-AMOUNT
   END-READ.
Avoid using PERFORM THRU statements, which can lead to maintenance issues if paragraphs are reordered or renamed. Use structured PERFORM statements with well-defined paragraph boundaries instead.
* Anti-pattern: Duplicating data definitions
* In Program A
01 CUSTOMER-RECORD.
   05 CUSTOMER-ID           PIC X(10).
   05 CUSTOMER-NAME         PIC X(30).
   05 CUSTOMER-ADDRESS.
      10 STREET-ADDRESS-1   PIC X(30).
      10 STREET-ADDRESS-2   PIC X(30).
      10 CITY               PIC X(20).
      10 STATE              PIC X(2).
      10 ZIP-CODE           PIC X(10).

* In Program B (duplicated definition)
01 CUSTOMER-RECORD.
   05 CUSTOMER-ID           PIC X(10).
   05 CUSTOMER-NAME         PIC X(30).
   05 CUSTOMER-ADDRESS.
      10 STREET-ADDRESS-1   PIC X(30).
      10 STREET-ADDRESS-2   PIC X(30).
      10 CITY               PIC X(20).
      10 STATE              PIC X(2).
      10 ZIP-CODE           PIC X(10).

* Better approach: Use COPY books
* In CUSTOMER.CPY file
01 CUSTOMER-RECORD.
   05 CUSTOMER-ID           PIC X(10).
   05 CUSTOMER-NAME         PIC X(30).
   05 CUSTOMER-ADDRESS.
      10 STREET-ADDRESS-1   PIC X(30).
      10 STREET-ADDRESS-2   PIC X(30).
      10 CITY               PIC X(20).
      10 STATE              PIC X(2).
      10 ZIP-CODE           PIC X(10).

* In Program A and Program B
COPY CUSTOMER.
Use COPY books to share common data definitions and code across multiple programs. This reduces duplication and ensures consistency.
* Anti-pattern: Cryptic variable names
01 X1              PIC 9(5)V99.
01 X2              PIC 9(5)V99.
01 X3              PIC 9(5)V99.

COMPUTE X3 = X1 * X2.

* Better approach: Descriptive variable names
01 ITEM-PRICE      PIC 9(5)V99.
01 QUANTITY        PIC 9(5)V99.
01 TOTAL-COST      PIC 9(5)V99.

COMPUTE TOTAL-COST = ITEM-PRICE * QUANTITY.
Use meaningful and descriptive names for variables, paragraphs, and sections. This makes your code more readable and easier to understand and maintain.
* Anti-pattern: Minimal error handling
READ CUSTOMER-FILE
   INVALID KEY DISPLAY "Error reading file"
END-READ.

* Better approach: Comprehensive error handling
READ CUSTOMER-FILE
   INVALID KEY
      MOVE "Y" TO ERROR-FOUND
      MOVE "Customer record not found" TO ERROR-MESSAGE
      PERFORM LOG-ERROR
      PERFORM DISPLAY-ERROR-TO-USER
END-READ.

IF ERROR-FOUND = "Y"
   PERFORM ERROR-RECOVERY-ROUTINE
END-IF.
Implement proper error handling with detailed error messages, logging, and recovery procedures. This helps diagnose and resolve issues more quickly.
* Anti-pattern: Inappropriate use of REDEFINES
01 MULTI-PURPOSE-FIELD.
   05 NUMERIC-VALUE         PIC 9(10).
   05 TEXT-VALUE REDEFINES NUMERIC-VALUE PIC X(10).
   05 DATE-VALUE REDEFINES NUMERIC-VALUE.
      10 YEAR               PIC 9(4).
      10 MONTH              PIC 9(2).
      10 DAY                PIC 9(2).
      10 FILLER             PIC 9(2).

* Better approach: Use separate fields for different purposes
01 CUSTOMER-DATA.
   05 CUSTOMER-ID           PIC 9(10).
   05 CUSTOMER-NAME         PIC X(30).
   05 REGISTRATION-DATE.
      10 REG-YEAR           PIC 9(4).
      10 REG-MONTH          PIC 9(2).
      10 REG-DAY            PIC 9(2).
Avoid using REDEFINES to repurpose the same memory location for different data types. This can lead to data corruption and maintenance issues. Use separate fields for different purposes instead.
* Anti-pattern: Unstructured code
PROCEDURE DIVISION.
MAIN-LOGIC.
   OPEN INPUT CUSTOMER-FILE.
   OPEN OUTPUT REPORT-FILE.
   READ CUSTOMER-FILE
      AT END MOVE "Y" TO END-OF-FILE-FLAG
   END-READ.
   PERFORM UNTIL END-OF-FILE-FLAG = "Y"
      MOVE CUSTOMER-NAME TO REPORT-NAME
      MOVE CUSTOMER-BALANCE TO REPORT-BALANCE
      WRITE REPORT-RECORD
      READ CUSTOMER-FILE
         AT END MOVE "Y" TO END-OF-FILE-FLAG
      END-READ
   END-PERFORM.
   CLOSE CUSTOMER-FILE.
   CLOSE REPORT-FILE.
   STOP RUN.

* Better approach: Structured code with modular design
PROCEDURE DIVISION.
MAIN-LOGIC.
   PERFORM INITIALIZATION.
   PERFORM PROCESS-RECORDS UNTIL END-OF-FILE.
   PERFORM TERMINATION.
   STOP RUN.

INITIALIZATION.
   OPEN INPUT CUSTOMER-FILE.
   OPEN OUTPUT REPORT-FILE.
   PERFORM READ-CUSTOMER-RECORD.

PROCESS-RECORDS.
   PERFORM WRITE-REPORT-RECORD.
   PERFORM READ-CUSTOMER-RECORD.

WRITE-REPORT-RECORD.
   MOVE CUSTOMER-NAME TO REPORT-NAME.
   MOVE CUSTOMER-BALANCE TO REPORT-BALANCE.
   WRITE REPORT-RECORD.

READ-CUSTOMER-RECORD.
   READ CUSTOMER-FILE
      AT END MOVE "Y" TO END-OF-FILE-FLAG
   END-READ.

TERMINATION.
   CLOSE CUSTOMER-FILE.
   CLOSE REPORT-FILE.
Use structured programming principles with modular design. Break down your program into logical, self-contained modules with clear responsibilities.
* Anti-pattern: Nested IF statements
IF TRANSACTION-TYPE = "S"
   IF CUSTOMER-TYPE = "R"
      IF AMOUNT > 1000
         COMPUTE DISCOUNT = AMOUNT * 0.10
      ELSE
         COMPUTE DISCOUNT = AMOUNT * 0.05
      END-IF
   ELSE IF CUSTOMER-TYPE = "P"
      IF AMOUNT > 1000
         COMPUTE DISCOUNT = AMOUNT * 0.15
      ELSE
         COMPUTE DISCOUNT = AMOUNT * 0.10
      END-IF
   END-IF
END-IF.

* Better approach: Use EVALUATE
EVALUATE TRUE
   WHEN TRANSACTION-TYPE = "S" AND CUSTOMER-TYPE = "R" AND AMOUNT > 1000
      COMPUTE DISCOUNT = AMOUNT * 0.10
   WHEN TRANSACTION-TYPE = "S" AND CUSTOMER-TYPE = "R"
      COMPUTE DISCOUNT = AMOUNT * 0.05
   WHEN TRANSACTION-TYPE = "S" AND CUSTOMER-TYPE = "P" AND AMOUNT > 1000
      COMPUTE DISCOUNT = AMOUNT * 0.15
   WHEN TRANSACTION-TYPE = "S" AND CUSTOMER-TYPE = "P"
      COMPUTE DISCOUNT = AMOUNT * 0.10
   WHEN OTHER
      MOVE ZEROS TO DISCOUNT
END-EVALUATE.
Use the EVALUATE statement for complex conditional logic instead of deeply nested IF statements. This makes your code more readable and easier to maintain.