SAS Program Usage Analysis in IBM Application Discovery

Our mainframe environment includes significant SAS batch processing for critical business functions (financial analytics, actuarial reporting, data transformations). Currently, IBM Application Discovery cannot analyze SAS programs to extract:

• Program-to-program dependencies (PROC calls, %INCLUDE statements)
• Data lineage (INPUT/OUTPUT datasets, DB2 table access via SAS/ACCESS)
• Macro usage and relationships
• Job-to-SAS-program execution flows
   
• Requested Functionality:

Enable IBM Application Discovery to parse SAS code and capture:

Program Structure:

• DATA steps and PROC steps
• Internal program calls (%INCLUDE, CALL EXECUTE)
• Macro definitions (%MACRO/%MEND) and invocations

Data Access Patterns:

• INFILE/DATALINES (file inputs)
• SET/MERGE/UPDATE (SAS dataset usage)
• PROC SQL (DB2, Oracle, etc. table access)
• LIBNAME statements (data source connections)

Dependencies:

• SAS programs calling other SAS programs
• JCL jobs executing SAS programs (EXEC SAS/EXEC SASTOOL)
• Cross-references between macros and programs
• Format libraries and user-defined functions

Call Graph Generation:

• Similar to existing COBOL CALL/PERFORM analysis
• Visualize SAS program hierarchies
• Track macro expansion and code generation

Example Code Pattern (for reference):

/* SAS programs use patterns like: */
%INCLUDE SASMACRO(UTILITY);      /* Similar to COBOL COPY */
CALL EXECUTE('%PROCESS_DATA');   /* Dynamic program call */
PROC SQL;                         /* DB2 table access */
  SELECT * FROM DB2SCHEMA.TABLE;
QUIT;