Gary Diehl wrote on 11/10/2006 11:05:27 AM: > I honestly don't know what the problems were with the records that > caused ICETOOL to abend and IFASMFDP to work fine. A manual dump of > them to a readable file using SAS didn't show anything glaringly unusual > on eyeball-inspection. > > When I ran IFASMFDP to "clean up" this production abend, it came back > and reported 0 records in error. So why did ICE find a problem and > abend while IFA copied it fine and didn't report an error?
Since you had SPANINC=RC16 and DFSORT terminated for an incomplete spanned record, I have no reason to doubt that there was an incomplete spanned record in the file. I don't know much about IFASMFDP. Maybe it's default is to handle incomplete spanned records automatically or maybe it doesn't detect them, in which case there could be incomplete spanned records in the file and you wouldn't know it. If you had used SPANINC=RC0 or SPANINC=RC0 with DFSORT, maybe DFSORT would have handled the incomplete spanned records as well and you wouldn't have noticed them. The fact that DFSORT reported a problem when you told it to report a problem (SPANINC=RC16) and that IFASMFDP didn't report a problem might just mean that IFASMFDP doesn't report such problems when it finds them or doesn't find them. Maybe somebody else who knows more about how IFASMFDP handles incomplete spanned records could shed some light on this. The fact is, since you were using SPANINC=RC16, you really don't know whether DFSORT would have handled the incomplete spanned records or not with SPANINC=RC0 or SPANINC=RC4. So we're comparing apples to oranges here. >I don't know, but I can tell you that it was a repeatable issue, and using IFA > to clean it up was a whole lot easier than trying to use SAS to isolate > and delete the record in error and try ICE again (and maybe have it > abend again and again, as each new "bad" record was found and had to be > manually worked around). But you told DFSORT to terminate with SPANINC=RC16. If you didn't want it to terminate, you should have used SPANINC=RC0 or SPANINC=RC4. But maybe you didn't know that SPANINC=RC16 was the default and could be changed? > I'm sure we must be running the default RC16 you mentioned. If it's > that noteworthy of pursuing, we could probably arrange a set of test > data to troubleshoot the nature of the failure. Well, it would be interesting to know if you still got a failure with SPANINC=RC0 or SPANINC=RC4. If so, then we'd know that IFASMFDP can handle or ignores bad spanned record situations that cause DFSORT to terminate. But we really don't know that based on what you were doing. Frank Yaeger - DFSORT Team (IBM) - [EMAIL PROTECTED] Specialties: PARSE, JFY, SQZ, ICETOOL, IFTHEN, OVERLAY, Symbols, Migration => DFSORT/MVS is on the Web at http://www.ibm.com/storage/dfsort/ ---------------------------------------------------------------------- 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
