Positions 23-46 (Record Identification Codes)

Entries in positions 23 through 46 identify each record type in the input file. One to three identification codes can be entered on each specification line. More than three record identification codes can be specified on additional lines with the AND/OR relationship. If the file contains only one record type, the identification codes can be left blank; however, a record identifying indicator entry (positions 21 and 22) and a sequence entry (positions 17 and 18) must be made.

Note: Record identification codes are not applicable for graphic or UCS-2 data type processing: record identification is done on single byte positions only.

Three sets of entries can be made in positions 23 through 46: 23 through 30, 31 through 38, and 39 through 46. Each set is divided into four groups: position, not, code part, and character.

The following table shows which categories use which positions in each set.

Category	23-30	31-38	39-46
Position	23-27	31-35	39-43
Not	28	36	44
Code Part	29	37	45
Character	30	38	46

Entries in these sets need not be in sequence. For example, an entry can be made in positions 31 through 38 without requiring an entry in positions 23 through 30. Entries for record identification codes are not necessary if input records within a file are of the same type. An input specification containing no record identification code defines the last record type for the file, thus allowing the handling of any record types that are undefined. If no record identification codes are satisfied, control passes to the RPG IVexception/error handling routine.