Re: Designing an extension for feature-space similarity search

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Jay Levitt <jay(dot)levitt(at)gmail(dot)com>
Cc: PG Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Designing an extension for feature-space similarity search
Date: 2012-02-16 19:34:34
Message-ID: 14004.1329420874@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Jay Levitt <jay(dot)levitt(at)gmail(dot)com> writes:
> Perfect. Composite types are exactly what I need here; the application can
> declare its composite type and provide distance functions for each member,
> and the extension can use those to calculate similarity. How do I introspect
> the composite type's pg_class to see what it contains? I assume there's a
> better way than SPI on system catalogs :)

Definitely. Take a look at record_out, record_cmp, and sibling
functions on generic composites (src/backend/utils/adt/rowtypes.c).
You might or might not feel like wiring in more assumptions than those
have about the possible contents of the record type.

regards, tom lane

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Marti Raudsepp 2012-02-16 19:49:43 Re: possible new option for wal_sync_method
Previous Message Dimitri Fontaine 2012-02-16 18:57:39 Re: [trivial patch] typo in doc/src/sgml/sepgsql.sgml