> I have managed to add support for the Oracle Array Interface in > DBD::Oracle. > The original code comes from an old patch that Kristian Nielsen had > created > a while back but never seem to get into the to trunk. > > I was hoping someone could do a benchmark type test on loading many > thousand > or millions or records, as this seem to be what people want to use the > interface for.
I'm very interested in this, but for a slightly different reason - I
don't have huge data volumes, just hundreds to maybe a few thousand rows
at a time, but in one particular process I am having to connect over a
medium latency (120ms) transatlantic network link. The network overheads
for single rows inserts are killing performance here.
Tried compiling on Cygwin to start with; this failed as an earlier change
in LOB support didn't add some lines to oci.def - patch attached to fix
it (this applies to both trunk and array branch).
Have also now compiled and tested on Solaris; results from a simple test
inserting batches of 1000 rows to a remote database:
Rate singlerow arrayinterface
singlerow 4.76/s -- -87%
arrayinterface 35.7/s 650% --
So that looks quite promising.
thanks,
--
Andy Hassall / <[EMAIL PROTECTED]> / <http://www.andyh.co.uk>
<http://www.andyhsoftware.co.uk/space> Space: disk usage analysis tool
dbd-oracle-oci.def.patch
Description: Binary data
