Re: 7.0 FK trigger question - Mailing list pgsql-hackers

From Adriaan Joubert
Subject Re: 7.0 FK trigger question
Date
Msg-id 38EACFB5.3F562696@albourne.com
Whole thread Raw
In response to Re: 7.0 FK trigger question  (wieck@debis.com (Jan Wieck))
Responses Re: 7.0 FK trigger question  (wieck@debis.com (Jan Wieck))
List pgsql-hackers
>

Thanks for the reply. For the time begin I've solved this by copying every
table in the database to a backup table without any constraints, recreating
the tables and copying the data back in. I have to be a bit careful with
doing it all in the right order, although I think I can solve this by doing
everything in a transaction as the constraints are only checked at the end of
transaction?

> >
> > So I guess my question really boils down to: is it possible to write a
> > function that drops a foreign key trigger or re-instates it? This should
> > really be ALTER TABLE  table ALTER COLUMN column (DROP|CREATE)
> > CONSTRAINT.... or something along those lines.
>
>     There's   still   something  missing  in  ALTER  TABLE.  DROP
>     CONSTRAINT is one of  them,  but  since  your  sequencs  with
>     renaming  the old etc. is the safest possibility anyway, it's
>     not that high priority.

OK, I'm definitely not being very bright here, but i cannot get my system to
accept the alter column commands. An example on the man pages ,ay help a lot
here! I tried

test=# create table t (i int4);
CREATE
test=# create table t1 (k int4);
CREATE
test=# alter table t1 alter column k add constraint references t(i);
ERROR:  parser: parse error at or near "add"
test=# alter table t1 alter column k constraint references t(i);
ERROR:  parser: parse error at or near "constraint"
test=# alter table t1 alter k constraint references t(i);
ERROR:  parser: parse error at or near "constraint"
test=# alter table t1 alter column k create constraint references t(i);
ERROR:  parser: parse error at or near "create"

So what am I doing wrong?

Thanks,

Adriaan



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: postgres crash on CURSORS
Next
From: Adriaan Joubert
Date:
Subject: Re: BIT datatype (Fixed)