>>For compatibility and migration purposes, is there a way for DFSORT to accept this specification?
Unfortunately NO. The other product had to introduce the formats for compensating the rich functionality of JNF1CNTL and JNF2CNTL in DFSORT. DFSORT gives you the entire INREC formatting with JNF1 and JNF2. >>I could check and see if Syncsort actually required this specification or if I did it by habit. If the key formats are of the same format (like in this case both key fields are character data) , you absolutely don't require the format. However if your key has ZD and PD combination then you can use JNF1/JNF2 to normalize one of the key to match the other. for example as 4 byte Packed decimal field is compared to 9 byte zoned decimal number JOINKEYS FILE=F1,FIELDS=(1,4,PD,A) JOINKEYS FILE=F2,FIELDS=(6,9,ZD,A) You can change this to JOINKEYS FILE=F1,FIELDS=(81,9,A) JOINKEYS FILE=F2,FIELDS=(6,9,A) //JNF1CNTL DD * INREC OVERLAY=(81:1,4,PD,ZD,LENGTH=9) $ convert 4 byte PD to 9 byte ZD format /* Thanks, Kolusu DFSORT Development IBM Corporation ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
