Re: Multifunction Indexes

From: Stephan Szabo <sszabo(at)megazone23(dot)bigpanda(dot)com>
To: <elein(at)varlena(dot)com>
Cc: David Wheeler <david(at)kineticode(dot)com>, <sfpug(at)postgresql(dot)org>
Subject: Re: Multifunction Indexes
Date: 2003-03-15 00:38:54
Message-ID: 20030314163257.Y87074-100000@megazone23.bigpanda.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: sfpug


On Fri, 14 Mar 2003, elein wrote:

> As a workaround, push the lower() function into the
> workflow() function. Of course if the workflow doesn't
> always want lower($1) then you'll have to overload or rename it.
>
> I'm don't know the structures like Stephen does.
> But if a plain expression parser were used consistently
> it should be able to enable expressions rather
> than single functions. The expression tree would need
> to be held rather than the function function pointer.
> And of course all of it should be immutable.

Yeah, right now IIRC the index structure basically effectively has:
Array of columns
Function oid for functional indexes

That's also why you can't currently have constants in them, etc...

It needs work, but it's also probably likely to break/touch alot of stuff
to change, so that's why it's probably not been done.

In response to

Browse sfpug by date

  From Date Subject
Next Message David Wheeler 2003-03-15 00:39:53 Re: Multifunction Indexes
Previous Message elein 2003-03-15 00:25:52 Re: Multifunction Indexes