On Wed, Jul 23, 2008 at 11:43 AM, Francisco Reyes
<lists@stringsutils.com> wrote:
> On 12:40 pm 07/23/08 Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> That's fairly hard to believe. Care to offer some details?
>
>
> I will dig that actual project and run explain analyze. Will likely not
> have it till middle of next week though because of a monthly process
> starting out Friday.
>
> However, I do have a current example where bitmap index scan was 3 times
> worse. On the extremely bad case the data set was 5 times larger than the
> sample below (1+ billion vs 215 million).
What is your work_mem set to?