Re: Why "ERROR: dtoi4: integer out of range" on pg_dump

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Martin Weinberg <weinberg(at)osprey(dot)astro(dot)umass(dot)edu>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Why "ERROR: dtoi4: integer out of range" on pg_dump
Date: 2001-09-03 21:46:29
Message-ID: 6676.999553589@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Martin Weinberg <weinberg(at)osprey(dot)astro(dot)umass(dot)edu> writes:
> DumpComment: SELECT failed: 'ERROR: dtoi4: integer out of range

Hmm. I can reproduce this error message if I suppose that you have
OIDs exceeding 2 billion. pg_dump will produce queries like:

regression=# select * from pg_description where objoid = 2500000000;
ERROR: dtoi4: integer out of range

A short-term workaround is to hack pg_dump so that it explicitly coerces
the literal to OID and/or quotes the literal:

regression=# select * from pg_description where objoid = 2500000000::oid;
objoid | classoid | objsubid | description
--------+----------+----------+-------------
(0 rows)

regression=# select * from pg_description where objoid = '2500000000';
objoid | classoid | objsubid | description
--------+----------+----------+-------------
(0 rows)

This is done in many places in pg_dump, but not in DumpComment which is
relatively new code :-(

A longer-term question is how to persuade the parser to get this right
without such help. I think that this is another variant of the
perennial numeric-precision issue and will not be real easy to fix.

regards, tom lane

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Oliver Elphick 2001-09-03 22:11:54 Re: Conditional operators ALL, ANY in WHERE clause
Previous Message Bruce Momjian 2001-09-03 21:19:44 Re: Re: Escaping strings for inclusion into SQL queries