I think it depends on how much data you are working with. REXX is very easy to use but costly to run. DFSORT will outperform REXX by orders of magnitude. You might be able to do all of your work with DFSORT and not need any REXX code at all. This would be the best.
John Reda Syncsort, Inc. -----Original Message----- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of McKown, John Sent: Friday, May 23, 2008 10:00 AM To: [email protected] Subject: Opinions? Subset input with DFSORT before processing in REXX? I write a lot of code in REXX. I also like Perl, but rarely use it on z/OS. Anyway, I was just trying to think (hard on the lead in to a 3 day weekend in the states). Would I gain very much if I subset my input file using DFSORT instead of subsetting in my REXX code. My example in this is the output from IRRDBU00. That creates a "flat file" unload of the RACF database, which I use to write a report. I use the 0200 records. In my shop they account for less than 5% of the records in the file. Seems to me that using DFSORT to subset the data, even if only used once, would result in less CPU, I/O, and elapsed time. What say ye? -- John McKown ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
