Re: postgresql 10.1 wrong plan in when using partitions bug

Mariel Cherkassky <mariel.cherkassky@gmail.com>

From: Mariel Cherkassky <mariel.cherkassky@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Andreas Kretschmer <andreas@a-kretschmer.de>, pgsql-performance@lists.postgresql.org
Date: 2018-02-04T15:42:23Z
Lists: pgsql-performance
Hi Tom,
Did you hear about any solution that is similar to oracle`s global index ?
Is there any way to query all the partitions with one index?

2018-02-04 17:39 GMT+02:00 Tom Lane <tgl@sss.pgh.pa.us>:

> Mariel Cherkassky <mariel.cherkassky@gmail.com> writes:
> > Great, it solved the issue. Seems problematic that the planner do full
> > scans on all partitions in the first case isnt it ? Seems like a bug ?
>
> to_date isn't an immutable function (it depends on timezone and possibly
> some other GUC settings).  So there's a limited amount that the planner
> can do with it.
>
>                         regards, tom lane
>