Re: How to insert a bulk of data with unique-violations very fast

On Sun, Jun 6, 2010 at 6:02 AM, Torsten Zühlsdorff
<foo@meisterderspiele.de> wrote:
> Scott Marlowe schrieb:
> Thank you very much for your example. Now i've got it :)
>
> I've test your example on a small set of my rows. While testing i've
> stumpled over a difference in sql-formulation. Using except seems to be a
> little slower than the more complex where not in (subquery) group by. Here
> is my example:

Yeah, to get a good idea you need a more realistic example.  Build
some tables with millions of rows using generate_series() and then
test against those.