Problem Deleting Referenced records

[Alex]

Hi,
I have two tables , A and B where table B has a foreign key constraint
to table A.

I want to delete all records in table A that are older than a certain
date that are not referenced by table B.

When I use a DELETE FROM the entire transaction fails as soon as a
referential integrity violation is detected.

Is that the normal behavior? Is there a way to allow the deletion to
complete but skipping all those records that are referenced ?

Thanks in advance for any help.

Alex

---------------------------(end of broadcast)---------------------------
TIP 7: don't forget to increase your free space map settings

[Bruno Wolff III]

On Mon, Nov 10, 2003 at 14:31:00 +0900,
Alex <al**@meerkatso ft.com> wrote:

Hi,
I have two tables , A and B where table B has a foreign key constraint
to table A.

I want to delete all records in table A that are older than a certain
date that are not referenced by table B.

When I use a DELETE FROM the entire transaction fails as soon as a
referential integrity violation is detected.

Is that the normal behavior? Is there a way to allow the deletion to
complete but skipping all those records that are referenced ?

I think you want to do something like:

delete from A where A.stamp < current_date - '1 month' and
not exists (select 1 from B where B.Aid = A.id);

It is also possible to join A and B on the delete command using a nonstandard
syntax, but not exists should be about as fast.

---------------------------(end of broadcast)---------------------------
TIP 5: Have you checked our extensive FAQ?

http://www.postgresql.org/docs/faqs/FAQ.html

[Alex]

Bruno,
thanks. I actually did it that way but having to join two tables each
1-2 million records makes this process rather time consuming.
I was hoping that the ON DELETE options in the constraint could handle
that.

It seems to be a bit odd that if I want to delete 100 records that are
not related to each other, and one record deletion fails that then the
entire delete process fails.

.... but I guess there is nothing I can do about.

Alex

Bruno Wolff III wrote:

On Mon, Nov 10, 2003 at 14:31:00 +0900,
Alex <al**@meerkatso ft.com> wrote:

Hi,
I have two tables , A and B where table B has a foreign key constraint
to table A.

I want to delete all records in table A that are older than a certain
date that are not referenced by table B.

When I use a DELETE FROM the entire transaction fails as soon as a
referential integrity violation is detected.

Is that the normal behavior? Is there a way to allow the deletion to
complete but skipping all those records that are referenced ?

I think you want to do something like:

delete from A where A.stamp < current_date - '1 month' and
not exists (select 1 from B where B.Aid = A.id);

It is also possible to join A and B on the delete command using a nonstandard
syntax, but not exists should be about as fast.

---------------------------(end of broadcast)---------------------------
TIP 5: Have you checked our extensive FAQ?

http://www.postgresql.org/docs/faqs/FAQ.html

---------------------------(end of broadcast)---------------------------
TIP 4: Don't 'kill -9' the postmaster

[Bruno Wolff III]

On Mon, Nov 10, 2003 at 16:20:21 +0900,
Alex <al**@meerkatso ft.com> wrote:

Bruno,
thanks. I actually did it that way but having to join two tables each
1-2 million records makes this process rather time consuming.
I was hoping that the ON DELETE options in the constraint could handle
that.
If only a small number of the 1-2 million records have old dates, than the
where not exists method might be faster. An index scan could be used
to find the records with old dates and then for each record an index
lookup could be done in table B to see if it should really be deleted.

It seems to be a bit odd that if I want to delete 100 records that are
not related to each other, and one record deletion fails that then the
entire delete process fails.

You can delete each record in its own transaction if you want that
behavior.

---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to ma*******@postg resql.org

[Alex]

Bruno,
I am not sure why but the whole delete proces with the where not exists
method took 3hrs, rather long I would say.

Table A had 2.5mil records
Table B had about 30k records

In addition Table C with about 2 mil records referenced a referenced A
(same key as B)
In both A and B about 150k records where deleted... but the process took
more than 3 hrs.

Its pretty long I would say. I noticed in the past that if I had
multiple foreign keys, referencing different tables like TableA <--
TableB <-- TableC then deletes are really slow... sometimes in the area
of one delete per second. Never really figured out why. (And yes I did
run a Vacuum or Vacuum analyze on the DB
or Tables).

Alex

Bruno Wolff III wrote:

On Mon, Nov 10, 2003 at 16:20:21 +0900,
Alex <al**@meerkatso ft.com> wrote:

Bruno,
thanks. I actually did it that way but having to join two tables each
1-2 million records makes this process rather time consuming.
I was hoping that the ON DELETE options in the constraint could handle
that.

If only a small number of the 1-2 million records have old dates, than the
where not exists method might be faster. An index scan could be used
to find the records with old dates and then for each record an index
lookup could be done in table B to see if it should really be deleted.

It seems to be a bit odd that if I want to delete 100 records that are
not related to each other, and one record deletion fails that then the
entire delete process fails.

You can delete each record in its own transaction if you want that
behavior.

---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to ma*******@postg resql.org

---------------------------(end of broadcast)---------------------------
TIP 4: Don't 'kill -9' the postmaster

[Bruno Wolff III]

On Wed, Nov 12, 2003 at 00:08:02 +0900,
Alex <al**@meerkatso ft.com> wrote:

Bruno,
I am not sure why but the whole delete proces with the where not exists
method took 3hrs, rather long I would say.

Table A had 2.5mil records
Table B had about 30k records

In addition Table C with about 2 mil records referenced a referenced A
(same key as B)
In both A and B about 150k records where deleted... but the process took
more than 3 hrs.

Its pretty long I would say. I noticed in the past that if I had
multiple foreign keys, referencing different tables like TableA <--
TableB <-- TableC then deletes are really slow... sometimes in the area
of one delete per second. Never really figured out why. (And yes I did
run a Vacuum or Vacuum analyze on the DB
or Tables).

If you do a lot of deleting of rows in referenced tables, then you probably
want to create indexes in the referencing tables for the key. This will
allow for quicker checks by the referential integrity checking triggers.

---------------------------(end of broadcast)---------------------------
TIP 7: don't forget to increase your free space map settings