Job wrote:
I am sure it isn't real easy to do, otherwise Tim would have
> done it long ago.
Tim responded:
Actually I've been moving the code in that direction for several releases. It's (possibly) most of the way there. Patches welcome.
This is good news. I'll take a look and see what sense I can make of it.
Job noted:
the performance differences are huge. As an example, I saw 60 seconds of database cpu time from inserting 10k rows (row by row or slow by slow) in C++ go to 2 seconds with an array insert instead.
Of course, for bulk data loading you can also just wrap SQL*Loader, which is what we do for things that fit that model.
-- Steve Sapovits [EMAIL PROTECTED]