Re: How to use int2vector

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Rich Cullingford <rculling(at)sysd(dot)com>
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: How to use int2vector
Date: 2004-01-29 17:11:51
Message-ID: 24238.1075396311@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

Rich Cullingford <rculling(at)sysd(dot)com> writes:
> In trying to write a utility to drop/restore table indexes (as an aid to
> a fast reload of the associated tables), I ran into the definition of
> the field pg_index.indkey as an int2vector.

> How can I deal with this type in plpgsql?

You should be able to index into it (indkey[0], etc), but I cannot think
of any good reason to be looking at the details of the pg_index row for
your purposes. It is certainly not reasonable to think of writing
plpgsql code to reconstruct 7.4 expressional index definitions from
what's in pg_index :-(. I would suggest using the pg_get_indexdef
function instead. Perhaps something like

regression=# select pg_get_indexdef(indexrelid) from pg_index where indrelid = 'tenk1'::regclass;
pg_get_indexdef
-----------------------------------------------------------
CREATE INDEX tenk1_hundred ON tenk1 USING btree (hundred)
CREATE INDEX tenk1_unique2 ON tenk1 USING btree (unique2)
CREATE INDEX tenk1_unique1 ON tenk1 USING btree (unique1)
(3 rows)

regression=#

pg_get_indexdef() exists at least as far back as 7.0, and probably
further (pg_dump has relied on it for a long time).

regards, tom lane

In response to

Browse pgsql-admin by date

  From Date Subject
Next Message Tom Lane 2004-01-29 17:15:31 Re: Error seen when vacuuming pg_largeobject table
Previous Message Chris White (cjwhite) 2004-01-29 17:03:35 Re: Error seen when vacuuming pg_largeobject table