Re: Indexing of LIKE queries is broken in current sources - Mailing list pgsql-hackers

From Thomas Lockhart
Subject Re: Indexing of LIKE queries is broken in current sources
Date
Msg-id 39C05C2D.C1271929@alumni.caltech.edu
Whole thread Raw
In response to Indexing of LIKE queries is broken in current sources  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: Indexing of LIKE queries is broken in current sources
List pgsql-hackers
> It seems that the parser now emits some kind of function call for LIKE
> expressions, whereas the optimizer's code to use indexes for LIKE is
> looking for an operator.
> I have more pressing things to do than try to teach the optimizer about
> looking for function calls as well as operators, so I recommend
> reverting the parser's output to what it was.

Oh, that's bad. I changed it to the function call to allow
three-parameter LIKE clauses, which is a perfectly reasonable thing to
do imho.

This LIKE shortcircuit stuff is a hack anyway, but where should I look
in the optimizer?
                      - Thomas


pgsql-hackers by date:

Previous
From: The Hermit Hacker
Date:
Subject: using a join in an 'INSERT ... SELECT' ...
Next
From: Stephan Szabo
Date:
Subject: Re: using a join in an 'INSERT ... SELECT' ...