Previously, I said: Two replies I received suggested that I copy the LRECL=32756 datasets to a smaller LRECL, and compare those. I didn't put much hope in this because I had already run the compare overriding both input LRECLs to be 29175, and I have had considerable success for many years overriding DCBs. So I was confident nothing would change.
Well, I'm happy to report that changing the LRECL of the datasets used in the compare worked! and I'm very happy. But now, I'm puzzled: why didn't the input file LRECL overrides work?! This just in: ... Anyway SuperC (in its wisdom) takes the input datasets characteristics (LRECL, BLKSIZE etc) and overrides them when the DELDD DD is opened. (Even if lower LRECL is defined on the DCB) ... This response is from IBM Support. The problem I was having was attempting to override the DCB of a dataset with a known highest LRECL less than the file's DSCB LRECL of 32756. The feature, UPDPDEL (Update Prefixed Delta lines) prefixes the delta data with info I found very convenient to exploit, as opposed to the non-prefixed method, UPDLDEL (Update Long Control with all matches and delta changes) that required coding I don't have time to write (yet). So, IBM has missed an opportunity to document their relatively nonstandard (but perfectly valid) program override of the DCB before OPEN. I feel that this is not a good thing when it's done in a utility! The doc does say: "The update data set is a variable-length data set reflecting the input source files' characteristics." But I think a little more info could help folks exploit the power of such an awesome facility as SUPERC. FYI Michael ---------------------------------------------------------------------- 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