Postgres Pro
Downloads
Home   >   mailing lists

Thread: last built-in oid

last built-in oid

From
"Christopher Kings-Lynne"
Date:
Hi,

What is the current way of getting the last built-in oid?

I looked at the source of pg_dump, and it does this:

SELECT datlastsysoid from pg_database where datname = 'dbname'

But as far as I can tell, the datlastsysoid field does not exist in
pg_database.

What gives?

--
Christopher Kings-Lynne
Family Health Network (ACN 089 639 243)

Re: last built-in oid

From
Philip Warner
Date:
At 11:27 24/11/00 +0800, Christopher Kings-Lynne wrote:
>
>SELECT datlastsysoid from pg_database where datname = 'dbname'
>
>But as far as I can tell, the datlastsysoid field does not exist in
>pg_database.
>

If you build from CVS and do an initdb, you will find datlastsysoid should
exist...


----------------------------------------------------------------
Philip Warner                    |     __---_____
Albatross Consulting Pty. Ltd.   |----/       -  \
(A.B.N. 75 008 659 498)          |          /(@)   ______---_
Tel: (+61) 0500 83 82 81         |                 _________  \
Fax: (+61) 0500 83 82 82         |                 ___________ |
Http://www.rhyme.com.au          |                /           \|                                |    --________--
PGP key available upon request,  |  /
and from pgp5.ai.mit.edu:11371   |/

RE: last built-in oid

From
"Christopher Kings-Lynne"
Date:
Ah.  What about in 7.0.3 and below?

Basically this because I am attempting to select all built-in functions.  I
would like it to be backwards compatible, if at all possible.

Chris

> -----Original Message-----
> From: Philip Warner [mailto:pjw@rhyme.com.au]
> Sent: Friday, November 24, 2000 11:47 AM
> To: Christopher Kings-Lynne; Pgsql-Hackers
> Subject: Re: [HACKERS] last built-in oid
>
>
> At 11:27 24/11/00 +0800, Christopher Kings-Lynne wrote:
> >
> >SELECT datlastsysoid from pg_database where datname = 'dbname'
> >
> >But as far as I can tell, the datlastsysoid field does not exist in
> >pg_database.
> >
>
> If you build from CVS and do an initdb, you will find datlastsysoid should
> exist...
>
>
> ----------------------------------------------------------------
> Philip Warner                    |     __---_____
> Albatross Consulting Pty. Ltd.   |----/       -  \
> (A.B.N. 75 008 659 498)          |          /(@)   ______---_
> Tel: (+61) 0500 83 82 81         |                 _________  \
> Fax: (+61) 0500 83 82 82         |                 ___________ |
> Http://www.rhyme.com.au          |                /           \|
>                                  |    --________--
> PGP key available upon request,  |  /
> and from pgp5.ai.mit.edu:11371   |/
>

Re: last built-in oid

From
Tom Lane
Date:
"Christopher Kings-Lynne" <chriskl@familyhealth.com.au> writes:
> Ah.  What about in 7.0.3 and below?

There is no good way --- if there were, we'd not have bothered to invent
datlastsysoid.  pg_dump used to use the OID of the template1 database
as an estimate of the last built-in OID.  This was wrong to begin with,
and is completely untenable in 7.1 (template1's OID is now 1).

> Basically this because I am attempting to select all built-in
> functions.

If you only care about functions then it's probably possible to
hard-wire an assumption that system functions have OIDs < 16384.
Right now all built-in functions have manually-assigned OIDs,
so that works.  But I wouldn't want to promise that it'll work
forever.  It already doesn't work for aggregates, for example
(were you including aggregates in "functions"?).
        regards, tom lane