MatterAI - Frontier LLM Intelligence at 70% Less Cost

ABAP Anti-Patterns Overview

ABAP, despite being a powerful language for SAP development, has several common anti-patterns that can lead to performance issues, maintainability problems, and bugs. Here are the most important anti-patterns to avoid when writing ABAP code.

SELECT * FROM Database Tables

" Anti-pattern: Using SELECT *
SELECT * FROM scarr INTO TABLE lt_carriers.

" Better approach: Select only needed fields
SELECT carrid carrname currcode
  FROM scarr
  INTO TABLE lt_carriers.

Avoid using SELECT * to retrieve all fields from database tables. This increases network traffic, memory usage, and processing time. Instead, explicitly list only the fields you need for your application logic.

Not Using WHERE Clause in Database Queries

" Anti-pattern: No WHERE clause, filtering in ABAP
SELECT carrid carrname currcode
  FROM scarr
  INTO TABLE lt_carriers.

LOOP AT lt_carriers INTO ls_carrier
  WHERE currcode = 'USD'.
  " Process USD carriers
ENDLOOP.

" Better approach: Filter at database level
SELECT carrid carrname currcode
  FROM scarr
  INTO TABLE lt_carriers
  WHERE currcode = 'USD'.

LOOP AT lt_carriers INTO ls_carrier.
  " Process USD carriers
ENDLOOP.

Always use appropriate WHERE clauses in your database queries to filter data at the database level. Retrieving all records and then filtering them in ABAP is inefficient, especially for large tables.

Using SELECT Inside Loops

" Anti-pattern: SELECT inside loop
SELECT carrid carrname
  FROM scarr
  INTO TABLE lt_carriers.

LOOP AT lt_carriers INTO ls_carrier.
  SELECT COUNT(*)
    FROM spfli
    WHERE carrid = ls_carrier-carrid
    INTO lv_count.
  
  ls_carrier-flight_count = lv_count.
  MODIFY lt_carriers FROM ls_carrier.
ENDLOOP.

" Better approach: Use JOIN or nested SELECT
SELECT s~carrid s~carrname COUNT(*) AS flight_count
  FROM scarr AS s
  LEFT OUTER JOIN spfli AS p ON s~carrid = p~carrid
  GROUP BY s~carrid s~carrname
  INTO TABLE lt_carriers.

Avoid executing database queries inside loops. This can lead to the “N+1 query problem” and severely impact performance. Instead, use JOINs, nested SELECTs, or FOR ALL ENTRIES to retrieve related data in a single query.

Not Using Internal Tables Effectively

" Anti-pattern: Inefficient internal table operations
DATA: lt_data TYPE TABLE OF ty_data,
      ls_data TYPE ty_data.

" Inefficient: Appending in a loop
DO 1000 TIMES.
  ls_data-id = sy-index.
  ls_data-value = 'Value'.
  APPEND ls_data TO lt_data.
ENDDO.

" Better approach: Use APPEND INITIAL LINE or table expressions
DO 1000 TIMES.
  APPEND INITIAL LINE TO lt_data ASSIGNING <fs_data>.
  <fs_data>-id = sy-index.
  <fs_data>-value = 'Value'.
ENDDO.

" Or even better in modern ABAP
lt_data = VALUE #( FOR i = 1 UNTIL i > 1000 ( id = i value = 'Value' ) ).

Use internal tables efficiently. Avoid operations like APPEND inside loops for large datasets. Instead, use APPEND INITIAL LINE with field symbols, table expressions, or modern ABAP constructs like VALUE operator for better performance.

Excessive Use of Global Variables

" Anti-pattern: Using global variables
REPORT zprogram.

DATA: gv_id TYPE i,
      gv_name TYPE string,
      gt_data TYPE TABLE OF ty_data.

START-OF-SELECTION.
  PERFORM get_data.
  PERFORM process_data.
  PERFORM display_results.

FORM get_data.
  SELECT * FROM ztable INTO TABLE gt_data.
ENDFORM.

FORM process_data.
  " Uses global gt_data
ENDFORM.

" Better approach: Pass parameters between procedures
REPORT zprogram.

START-OF-SELECTION.
  DATA: lt_data TYPE TABLE OF ty_data.
  
  PERFORM get_data CHANGING lt_data.
  PERFORM process_data USING lt_data.
  PERFORM display_results USING lt_data.

FORM get_data CHANGING pt_data TYPE ty_table.
  SELECT * FROM ztable INTO TABLE pt_data.
ENDFORM.

FORM process_data USING pt_data TYPE ty_table.
  " Uses parameter pt_data
ENDFORM.

Avoid excessive use of global variables. They create hidden dependencies between procedures, make code harder to test and maintain, and can lead to unexpected behavior. Instead, pass parameters between procedures and use local variables when possible.

Not Using ABAP Objects

" Anti-pattern: Procedural programming for complex logic
REPORT zprogram.

DATA: gt_data TYPE TABLE OF ty_data.

START-OF-SELECTION.
  PERFORM get_data.
  PERFORM process_data.
  PERFORM display_results.

" Many FORM routines with complex logic

" Better approach: Use ABAP Objects
REPORT zprogram.

START-OF-SELECTION.
  DATA(lo_data_processor) = NEW zcl_data_processor( ).
  
  lo_data_processor->get_data( ).
  lo_data_processor->process( ).
  lo_data_processor->display_results( ).

" Class implementation
CLASS zcl_data_processor DEFINITION.
  PUBLIC SECTION.
    METHODS: constructor,
             get_data,
             process,
             display_results.
  
  PRIVATE SECTION.
    DATA: mt_data TYPE TABLE OF ty_data.
ENDCLASS.

CLASS zcl_data_processor IMPLEMENTATION.
  " Method implementations
ENDCLASS.

Use ABAP Objects (classes and interfaces) for complex applications instead of procedural programming with FORM routines. Object-oriented programming provides better encapsulation, reusability, and maintainability.

Hardcoding Values

" Anti-pattern: Hardcoding values
SELECT * FROM sflight
  INTO TABLE lt_flights
  WHERE carrid = 'LH'
    AND connid = '0400'.

IF sy-subrc = 0.
  " Process flights
ENDIF.

" Better approach: Use constants and configuration
CONSTANTS: lc_carrier_id TYPE s_carr_id VALUE 'LH',
           lc_connection_id TYPE s_conn_id VALUE '0400'.

" Even better: Use customizing tables
SELECT single carrier connection
  FROM zconfig
  INTO (lv_carrier, lv_connection)
  WHERE config_id = 'DEFAULT'.

SELECT * FROM sflight
  INTO TABLE lt_flights
  WHERE carrid = lv_carrier
    AND connid = lv_connection.

Avoid hardcoding values in your code. Use constants, configuration tables, or customizing settings instead. This makes your code more maintainable and adaptable to different environments or requirements.

Not Using Proper Exception Handling

" Anti-pattern: No exception handling
METHOD create_order.
  DATA: lo_order TYPE REF TO cl_order.
  
  CREATE OBJECT lo_order.
  lo_order->set_customer( iv_customer ).
  lo_order->add_items( it_items ).
  lo_order->save( ).
ENDMETHOD.

" Better approach: Use exception handling
METHOD create_order.
  DATA: lo_order TYPE REF TO cl_order.
  
  TRY.
      CREATE OBJECT lo_order.
      lo_order->set_customer( iv_customer ).
      lo_order->add_items( it_items ).
      lo_order->save( ).
    CATCH cx_order_creation INTO DATA(lx_creation).
      RAISE EXCEPTION TYPE cx_process_error
        EXPORTING
          previous = lx_creation
          textid   = cx_process_error=>order_creation_failed.
    CATCH cx_system_error INTO DATA(lx_system).
      " Log system error
      RAISE RESUMABLE EXCEPTION TYPE cx_process_error
        EXPORTING
          previous = lx_system.
  ENDTRY.
ENDMETHOD.

Use proper exception handling in your code. Catch specific exceptions rather than generic ones, and handle them appropriately. Use class-based exceptions (CX_* classes) for better error information and propagation.

Not Using Modern ABAP Syntax

" Anti-pattern: Old ABAP syntax
DATA: lv_value TYPE string,
      lt_table TYPE TABLE OF ty_data,
      ls_data TYPE ty_data.

SELECT * FROM ztable INTO TABLE lt_table.

LOOP AT lt_table INTO ls_data WHERE id > 100.
  CONCATENATE 'Value:' ls_data-value INTO lv_value.
  WRITE: / lv_value.
ENDLOOP.

" Better approach: Use modern ABAP syntax
DATA(lt_table) = VALUE ty_table( ).

SELECT * FROM ztable INTO TABLE @lt_table.

LOOP AT lt_table ASSIGNING FIELD-SYMBOL(<ls_data>) WHERE id > 100.
  DATA(lv_value) = |Value: { <ls_data>-value }|.
  WRITE: / lv_value.
ENDLOOP.

Use modern ABAP syntax features like inline declarations, string templates, table expressions, and field symbols. Modern syntax is often more concise, readable, and can improve performance.

Inefficient String Handling

" Anti-pattern: Inefficient string concatenation
DATA: lv_result TYPE string.

DO 100 TIMES.
  CONCATENATE lv_result 'Item' sy-index INTO lv_result.
ENDDO.

" Better approach: Use string templates and efficient concatenation
DATA: lv_result TYPE string,
      lt_parts TYPE TABLE OF string.

DO 100 TIMES.
  APPEND |Item{ sy-index }| TO lt_parts.
ENDDO.

CONCATENATE LINES OF lt_parts INTO lv_result.

" Or use string templates directly
lv_result = |Item1Item2Item3...|.

Use efficient string handling techniques. Avoid repeated CONCATENATE operations in loops. Instead, collect strings in a table and concatenate them at once, or use string templates (|…|) for simple concatenations.

Not Using Code Inspector and ATC

" Anti-pattern: Not checking code quality
" Writing code without static analysis

" Better approach: Use Code Inspector and ATC
" Run Code Inspector checks regularly
" Set up ATC (ABAP Test Cockpit) for continuous code quality checks

Always use Code Inspector and ABAP Test Cockpit (ATC) to check your code for performance issues, security vulnerabilities, and adherence to coding standards. These tools can identify many common anti-patterns automatically.

Not Using Proper Authorization Checks

" Anti-pattern: Missing or improper authorization checks
METHOD display_employee_data.
  SELECT * FROM zemployees
    INTO TABLE @DATA(lt_employees).
  
  " Display sensitive data without authorization check
ENDMETHOD.

" Better approach: Implement proper authorization checks
METHOD display_employee_data.
  AUTHORITY-CHECK OBJECT 'ZEMPLOYEE'
    ID 'ACTVT' FIELD '03'.
  
  IF sy-subrc <> 0.
    RAISE EXCEPTION TYPE cx_no_authority.
  ENDIF.
  
  SELECT * FROM zemployees
    INTO TABLE @DATA(lt_employees).
  
  " Display data
ENDMETHOD.

Always implement proper authorization checks in your code, especially when accessing sensitive data or performing critical operations. Use AUTHORITY-CHECK statements and check the results before proceeding.

Not Using Proper Modularization

" Anti-pattern: Large, monolithic methods
METHOD process_sales_order.
  " 200+ lines of code doing many different things:
  " - Validating input
  " - Reading master data
  " - Calculating prices
  " - Creating order
  " - Updating inventory
  " - Sending notifications
ENDMETHOD.

" Better approach: Proper modularization
METHOD process_sales_order.
  validate_input( is_order_data ).
  
  DATA(lt_master_data) = read_master_data( is_order_data ).
  
  DATA(ls_pricing) = calculate_pricing(
    is_order_data = is_order_data
    it_master_data = lt_master_data
  ).
  
  DATA(lv_order_id) = create_order(
    is_order_data = is_order_data
    is_pricing = ls_pricing
  ).
  
  update_inventory( is_order_data ).
  
  send_notifications( lv_order_id ).
ENDMETHOD.

Break down large, monolithic methods into smaller, focused methods with clear responsibilities. Each method should do one thing well. This improves readability, maintainability, and testability of your code.

Not Using SAP Standard Code

" Anti-pattern: Reinventing the wheel
METHOD convert_date.
  DATA: lv_day TYPE c LENGTH 2,
        lv_month TYPE c LENGTH 2,
        lv_year TYPE c LENGTH 4,
        lv_result TYPE string.
  
  lv_day = iv_date+6(2).
  lv_month = iv_date+4(2).
  lv_year = iv_date(4).
  
  CONCATENATE lv_day '.' lv_month '.' lv_year INTO lv_result.
  
  rv_formatted_date = lv_result.
ENDMETHOD.

" Better approach: Use SAP standard functionality
METHOD convert_date.
  CALL FUNCTION 'CONVERT_DATE_TO_EXTERNAL'
    EXPORTING
      date_internal            = iv_date
    IMPORTING
      date_external            = rv_formatted_date
    EXCEPTIONS
      date_internal_is_invalid = 1
      OTHERS                   = 2.
  
  IF sy-subrc <> 0.
    RAISE EXCEPTION TYPE cx_invalid_date.
  ENDIF.
ENDMETHOD.

Leverage SAP standard functionality instead of reinventing the wheel. SAP provides many standard function modules, classes, and APIs for common tasks. Using them ensures consistency, reduces bugs, and saves development time.

Not Using Proper Naming Conventions

" Anti-pattern: Poor naming conventions
DATA: x TYPE i,
      y TYPE string,
      z TYPE TABLE OF ty_data.

METHOD m1.
  " Method implementation
ENDMETHOD.

" Better approach: Use descriptive names and conventions
DATA: lv_counter TYPE i,
      lv_employee_name TYPE string,
      lt_sales_data TYPE TABLE OF ty_sales_data.

METHOD calculate_sales_statistics.
  " Method implementation
ENDMETHOD.

Follow proper naming conventions for variables, methods, and classes. Use prefixes to indicate variable scope and type (e.g., lv_ for local variables, lt_ for local tables). Use descriptive names that clearly indicate the purpose of the entity.

Not Using ABAP Unit Tests

" Anti-pattern: No unit tests
CLASS zcl_calculator DEFINITION.
  PUBLIC SECTION.
    METHODS: add IMPORTING iv_a TYPE i iv_b TYPE i RETURNING VALUE(rv_result) TYPE i.
ENDCLASS.

CLASS zcl_calculator IMPLEMENTATION.
  METHOD add.
    rv_result = iv_a + iv_b.
  ENDMETHOD.
ENDCLASS.

" Better approach: Include unit tests
CLASS zcl_calculator_test DEFINITION FOR TESTING
  DURATION SHORT
  RISK LEVEL HARMLESS.
  
  PRIVATE SECTION.
    DATA: mo_cut TYPE REF TO zcl_calculator.  " Class Under Test
    
    METHODS:
      setup,
      test_add FOR TESTING.
ENDCLASS.

CLASS zcl_calculator_test IMPLEMENTATION.
  METHOD setup.
    CREATE OBJECT mo_cut.
  ENDMETHOD.
  
  METHOD test_add.
    DATA(lv_result) = mo_cut->add( iv_a = 2 iv_b = 3 ).
    
    cl_abap_unit_assert=>assert_equals(
      act = lv_result
      exp = 5
      msg = 'Addition calculation is wrong'
    ).
  ENDMETHOD.
ENDCLASS.

Write ABAP Unit tests for your code. Unit tests help catch bugs early, document expected behavior, and make it safer to refactor code. Follow test-driven development (TDD) principles when possible.

Not Considering Performance in Loops

" Anti-pattern: Inefficient operations in loops
LOOP AT lt_data INTO ls_data.
  CLEAR lt_temp.
  
  " Inefficient: Database access in loop
  SELECT * FROM ztable
    INTO TABLE lt_temp
    WHERE id = ls_data-id.
  
  " Inefficient: Nested loops
  LOOP AT lt_temp INTO ls_temp.
    " Process data
  ENDLOOP.
ENDLOOP.

" Better approach: Optimize loop operations
" Get all data at once before the loop
SELECT * FROM ztable
  INTO TABLE @DATA(lt_all_temp)
  FOR ALL ENTRIES IN @lt_data
  WHERE id = @lt_data-id.

" Use efficient loops
LOOP AT lt_data INTO ls_data.
  LOOP AT lt_all_temp INTO ls_temp WHERE id = ls_data-id.
    " Process data
  ENDLOOP.
ENDLOOP.

Be mindful of performance in loops. Avoid database access, screen operations, and complex calculations inside loops. Pre-fetch data, use efficient internal table operations, and consider alternatives to nested loops.

Not Using Data Dictionary Objects

" Anti-pattern: Not using data dictionary objects
TYPES: BEGIN OF ty_employee,
         id TYPE c LENGTH 10,
         name TYPE c LENGTH 40,
         department TYPE c LENGTH 20,
       END OF ty_employee.

DATA: lt_employees TYPE TABLE OF ty_employee.

" Better approach: Use data dictionary objects
DATA: lt_employees TYPE TABLE OF zemployee.

" Or with modern syntax
SELECT * FROM zemployee INTO TABLE @DATA(lt_employees).

Use data dictionary objects (tables, structures, data elements) instead of defining types locally. This ensures consistency across applications, enables central changes, and leverages SAP’s metadata capabilities.

Getting Started

AI Coding Agents

Axon Models

AI Code Reviews

Integrations

ABAP