The "demand" subroutines described in the previous examples can be generalized and used for any file.

Consider a "generalized" FETCH command like this:

     FETCH   FIELDS(#FIELD1 #FIELD2 ... #FIELDn) FROM_FILE(FILE) 
             WITH_KEY(#KEY1 ... #KEYn)

This can be replaced by executing a "demand" subroutine that exactly emulates the FETCH command like this:

     EXECUTE SUBROUTINE(GET_FILE) WITH_PARMS(#KEY1 ... #KEYn)

The actual "demand" subroutine in generalized format would look like this:

     SUBROUTINE NAME(GET_FILE) PARMS((#GETKEY1 *RECEIVED)
                                         "         "
                                     (#GETKEYn *RECEIVED))
 
     DEFINE FIELD(#GETKEY1) REFFLD(#KEY1)
        "     "      "           "
     DEFINE FIELD(#GETKEYn) REFFLD(#KEYn)
    
     DEF_LIST   NAME(#FILE) 
                FIELDS(#KEY1 ... #KEYn #FIELD1 ... #FIELDn)
                TYPE(*WORKING) ENTRYS(as required)
 
     LOC_ENTRY  IN_LIST(#FILE) WHERE('(#KEY1 = #GETKEY1) *AND
                                         "         "       "
                                      (#KEYn = #GETKEYn)')
 
     IF_STATUS  IS_NOT(*OKAY)
     FETCH   FIELDS(#FIELD1 #FIELD2 ... #FIELDn) FROM_FILE(FILE)
             WITH_KEY(#GETKEY1 ... #GETKEYn)
     ADD_ENTRY  TO_LIST(#FILE)
     ENDIF

Points to Note:

• This routine does not account for 2 possibilities.
• The first is that the working list may overflow. This could be solved by using the COUNTER parameter on the DEF_LIST command. Before the ADD_ENTRY command the value could be checked and the list cleared (CLR_LIST command) if greater than or equal to the ENTRYS value.
• The second is the fact that the requested record may not be found in either the working list or the database. This could be solved by an IF_STATUS after the FETCH command. If found the record would be added to the working list. If not found appropriate action could be taken (e.g.: ABORT command).