Hitoshi Harada <umi.tanuki@gmail.com> writes:
> 2011/4/27 Tom Lane <tgl@sss.pgh.pa.us>:
>> Joel Reymont <joelr1@gmail.com> writes:
>>> That's a 30x speedup, from 12 minutes down to 38s. Thanks Tom!
>> Huh, I would've bet on a lot more actually. �The nodeFunctionscan and
>> nodeAgg code must not be as inefficient as it looks on the surface ...
> Did you mean in that case you can optimize it by collapsing those
> nodes into one?
No, just that the overhead of that code looks like it ought to be much
more than the cost of a few multiplications ...
regards, tom lane