Rule Definition
String processing is expensive in EGL COBOL Gen for two reasons - All text data has to be converted to Unicode - EGL has to create/manage strings in its own heap memory manager and garbage collection. DBCHAR is for Double-Byte languages so doubt Halkbank will do this. UNICODE should only be used if intentionally trying to do Unicode processing in COBOL Gen as it is a fixed datatype. If not doing Unicode operations in EGL COBOL Gen, then should use CHAR...Unicode to/from CHAR would require data conversion.

Remediation
Try to reduce the use of Unicode or dbCHar datatypes.

Violation Code Sample

abc unicode(5);   // VIOLATION
def dbCHar(10);  // VIOLATION

EGL Best Practices: Coding For Performance. November 10, 2009. Edition 2