Thanks for the thanks !
Generally, when grouping stuff together, it is a good idea to have two
sorted lists, and to scan them simultaneously. I have already used this
solution several times outside of Postgres, and it worked very well (it
was with Berkeley DB and there were 3 lists to scan in order). The fact
that Python can very easily virtualize these lists using generators makes
it possible to do it without consuming too much memory.
> Pierre-Frederic, Paul,
>
> Thanks for your fast response (especially for the python code and
> performance figure) - I'll chase this up as a solution - looks most
> promising!
>
> Cheers,
>
> Damien
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org
>
