On Fri, 7 Nov 2014, Tom Lane wrote:
> That's an utterly insane number of partitions.
> I'm unsurprised.
>
> The inheritance infrastructure is only designed to support maybe 100
> child tables. If we ever get around to implementing a hard-wired
> partitioning scheme, it's possible that multiple thousands of partitions
> could be practical, but it isn't today. Sorry.
Okay, thanks.
I knew that it is not designed for such number of partitions, and that's
why I'm not using any of the CHECK() infrastructure and rather use the
inheritance to ensure the schema consistency of all the partitions. I was
just hoping that occasional queries from from parent would work (without
much worries about performance). I know that it did work in my tests when
the dataset was smaller (but with still the same number of children
tables), and was suprised that it blows so impressively even if the
actual data retrieved is zero. But fair enough, I won't do those queries
anymore.
Regards,
Sergey
*****************************************************
Sergey E. Koposov, PhD, Senior Research Associate
Institute of Astronomy, University of Cambridge
Madingley road, CB3 0HA, Cambridge, UK
Tel: +44-1223-337-551 Web: http://www.ast.cam.ac.uk/~koposov/