The reason I asked the question is that the Pro*C program appears to be writing the output to a file faster than my Perl program can write to the named pipe. The other end of the named pipe is a native database utility similar to the sqlldr (nzload). I don't know where the bottle neck is so I might have to break up the program into to separate parts. Have Perl dump the table to a file then feed the file using nzload and calculate the total time for each step. Peter
-----Original Message----- From: Garrett, Philip (MAN-Corporate) [mailto:[EMAIL PROTECTED] Sent: Thursday, August 31, 2006 2:44 PM To: DBI-Users Subject: RE: $dbh->{RowCacheSize} = 1000 Loo, Peter # PHX wrote: > > "RowCacheSize" (integer) > > What would you recommend I set this option to if I am sending all the > SELECTed rows to a named pipe? I'd recommend you leave it alone, unless your database roundtrips are really expensive. And if they are, the only thing that can tell you what the best value would be is to benchmark different settings. It will only help if the program on the other end of the named pipe can accept data faster than you can send it. Philip This E-mail message is for the sole use of the intended recipient(s) and may contain confidential and privileged information. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply E-mail, and destroy all copies of the original message.