Performance issue with using library QTEMP?

[mtharp]

I currently have a program that reads a file in arrival sequence. The data that is read in is the written to another arrival sequence file in QTEMP. The job runs much slower if QTEMP is used instead of some other production library. Any ideas? Thanks in advance for your help.

[David Abramowitz]

For sequential reads (not by key), use the "Number of REcords" parameter, and the Sequential(*YES nnn) parameter, on the OVRDBF command, where nnn = the number of records read at once. One guideline is that nnn * recodlength should be as close to 64k as possible, but under 64k. Dave