On Fri, Oct 3, 2008 at 1:54 AM, Greg Monroe <[EMAIL PROTECTED]> wrote:
> If we don't want to lose this functionality, we still need to do some > checking as to what the underlying column type is. I'm thinking that > this information should get added to the ColumnMap object and the > initialization routines changed to make the call to get the Meta > data. This would cache the meta data, cut the meta data requests to > just a few, and still maintain the underlying Village flexibility. When using the old PG driver there are a small number of DB queries to get type name information for the built-in PG types - and the results are cached. In the newer driver the cache is initialized with type names for these built-in types, so the handful of type name queries I was seeing in the old driver are completely gone when using the newer driver. So I didn't need to comment out the calls that get column type name information as these were not causing my "problem". See my previous message for the two lines that were generating all the extra DB queries. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
