> If I create a table with a datetime field with a default of 'now',
> every insert the value is the time of table creation instead of the
> time of insert, which is how it behaved in previous releases (I think
> this was even documented).
I can't recall it ever working that way, though before we discovered
that it didn't we all assumed that it _did_ work that way :)
The workaround is to define it as
... default datetime('now'::text)
which forces the string to be evaluated at runtime. The SQL symbol
CURRENT_TIMESTAMP also misbehaves in "default" clauses, and I'm
considering changing it a bit to get around the problem.
- Tom
In response to
Responses
pgsql-hackers by date
| Next: | From: Thomas G. Lockhart | Date: 1998-03-16 06:04:11 |
| Subject: Re: [HACKERS] Small changes for the "no excuses" release |
| Previous: | From: Bruce Momjian | Date: 1998-03-16 05:58:48 |
| Subject: Re: [HACKERS] initdb and xpg_user |