On 12/18/2012 01:24 PM, Tom Lane wrote:
> Bernhard Schrader <bernhard.schrader@innogames.de> writes:
>> Beside of that, we tested a little bit more with the failing query:
>> The statement which is causing the error is a big UPDATE-statement with
>> FROM. After some testing we figured out that the subselect in the
>> FROM-clause is working fine. And if we simplify the UPDATE-statement
>> it's also working. We're able to show the data and we're able to do
>> simple updates on the table. But the two things combined are not
>> working.
> Does the table being updated have any indexes on enum columns? I'm
> suspicious that the bogus OID is in an index page somewhere, and not
> in the table at all.
>
> If that is the answer, then reindexing said index would get rid of
> the problem (as well as all evidence that would help us find out how
> it happened ...)
>
>
Unless they can make a physical backup of the datadir first.
cheers
andrew