Re: Partial Indices vs. mixing columns and functions - Mailing list pgsql-general

From Tom Lane
Subject Re: Partial Indices vs. mixing columns and functions
Date
Msg-id 24984.995382693@sss.pgh.pa.us
Whole thread Raw
In response to Re: Partial Indices vs. mixing columns and functions  (Hiroshi Inoue <Inoue@tpf.co.jp>)
List pgsql-general
Hiroshi Inoue <Inoue@tpf.co.jp> writes:
> Both the functions used in functional indexes and the
> predicates used in partial indexes must be deterministic.
> Are users responsible for it ?
>>
>> Yes.  I don't see how the system could enforce that, without
>> prohibiting lots of useful cases.

> Shouldn't the functions be *cacheable* at least ?

Hmm.  We could do that, although of course it's not a bulletproof check.

Any objections out there?

            regards, tom lane

pgsql-general by date:

Previous
From: Tom Lane
Date:
Subject: Re: Partial Indices vs. mixing columns and functions
Next
From: "Ben-Nes Michael"
Date:
Subject: Re: VACUUM ANALYZE