>Unfortunately, the overhead of allocate/open/close/free for each member >does not yield good performance if that matters for your application.
In a previous project I had an image of a library on a Linux (my testbed) directory. I read that directory and processed each file separately. I used Rexx on Linux and it worked something like 15 minutes on 30000 files. But our target machine was IBM AIX machine (the fastedt machine possible according to the client,) so when I moved the process there it would take about 4 hours (I kid you not - and I had to process 4 libraries like that.) The solution was to move the thing back to the mainframe (which was considered to be a production machine as well.) It was still painfull because of the allocate/open/close/free mentioned above, but it took only an hour ZA ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN