Re: Implied Functional Index use - Mailing list pgsql-hackers

From Zeugswetter Andreas DCP SD
Subject Re: Implied Functional Index use
Date
Msg-id E1539E0ED7043848906A8FF995BDA579012B3EE1@m0143.s-mxs.net
Whole thread Raw
In response to Re: Implied Functional Index use  (Simon Riggs <simon@2ndquadrant.com>)
List pgsql-hackers
> There is a problem of implication here, AFAICS:
> When a user SQL asks
>     WHERE col1 = 7
> which equality level is meant when several exist?

Well, the operator must be unique, so there is no problem.
Unique in the sense that an operator with the same name ('=' in this
case)
and argument types cannot exist for more than one level of equality.
(and the level should not have an effect on the resolution)

So, when we see col1 = 7 we lookup the equality level of the operator
and decide whether it is strict enough for the particular optimization.

Andreas


pgsql-hackers by date:

Previous
From: Tatsuo Ishii
Date:
Subject: Re: [PERFORM] Sun Donated a Sun Fire T2000 to the PostgreSQL
Next
From: Dave Cramer
Date:
Subject: Re: postgresql.conf basic analysis tool