On Mon, 3 Nov 2008 19:27:14 -0600, John McKown wrote: >OK, the straight forward way to do this is to put both reports on disk, >using SDSF and its XDC command or equivalent. > >But, suppose these two reports are really, really, really large. Say 100 >million lines each. Wouldn't it be nice to be able to compare them >directly on the SPOOL? I know the first thing to look at is using SDSF to >compare the number of lines and byte count. If that doesn't agree, then >there is obviously something different. But, then again, what? > Batch SDSF allows you to PRINT to a DDNAME. That might be ALLOCATEd to a named pipe, with cksum(1) (or better; md5, sha, ...) on the other end.
Or you could fork(); run an SDSF in each process, and pipe both outputs ito diff(1). Long ago, WJS posted an interactive interface to batch SDSF on the Tools & Toys page. I judged it rudimentary; I could make major improvements. But this would be a derivative work, in violation of the explicit copyright. It would be worse if I distributed such a derivative work; worse yet if I got caught. While WJS might be flexible, the copyright belongs to IBM. The SDSF API might facilitate all this. Might one run two independent SDSFs under two TCBs? If not, two PIDs would more likely work. -- gil ---------------------------------------------------------------------- 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
