2010/7/29 Josh Berkus <josh@agliodbs.com>:
>
>> Or maybe checking against the source code and its documentation, if any.
>
> No, not really. What you really want to know is: what's the real
> planner overhead of having dozens/hundreds of partial indexes? What's
> the write overhead? There's no way you can derive that from the source
> code faster than you can test it.
Again, as the test would be rather killing for my group at this stage.
I think that knowing whether certain parts have been implemented
with linear or sub-linear (or whatever else) algorithms would
give good insights about scalability.
At a first glance it seems that for inheritance some bottleneck is
hindering a full exploit for table partitioning.
Is there anyone who knows whether those algorithms are linear or not?
And of course, I agree that real tests on real data will provide the real thing.
--
Vincenzo Romano at NotOrAnd Information Technologies
Software Hardware Networking Training Support Security
--
NON QVIETIS MARIBVS NAVTA PERITVS