| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> | 
|---|---|
| To: | "Gerard Krupa" <Gerard(dot)Krupa(at)marconi(dot)com> | 
| Cc: | pgsql-bugs(at)postgresql(dot)org | 
| Subject: | Re: pg_dump table ordering bug [8.0.1] | 
| Date: | 2005-02-10 15:29:54 | 
| Message-ID: | 26902.1108049394@sss.pgh.pa.us | 
| Views: | Whole Thread | Raw Message | Download mbox | Resend email | 
| Thread: | |
| Lists: | pgsql-bugs | 
"Gerard Krupa" <Gerard(dot)Krupa(at)marconi(dot)com> writes:
> I've experienced a similar problem with a data-only dump using 8.0.1's
> pg_dump.  It seems that the COPYs (or INSERTs) are generated in
> alphabetical order of table name, ignoring foreign key contraints and
> requiring the output to be post-processed.  In a full dump of the database
> the constraints are added after the data is inserted so there are no
> problems.  See the two attached dumps of the same database as an example.
This is not a bug.  In a data-only dump pg_dump cannot hope to know what
FK constraints may be applied to the data at load time --- there's no
particularly good reason to assume they are the same as what was in the
source database.  So it's your responsibility to order the loading steps
properly.  See pg_restore's -l/-L switches.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Luciano Almeida | 2005-02-10 19:33:42 | BUG #1475: error to install language pltcl | 
| Previous Message | Tom Lane | 2005-02-10 15:08:14 | Re: BUG #1473: Backend bus error, possibly due to ANALYZE |