From: | Dave Cramer <pg(at)fastcrypt(dot)com> |
---|---|
To: | Kris Jurka <books(at)ejurka(dot)com> |
Cc: | Mike Clements <mclements(at)actional(dot)com>, pgsql-jdbc(at)postgresql(dot)org |
Subject: | Re: Batch with keygen? |
Date: | 2005-11-08 00:01:04 |
Message-ID: | 1967C9F5-1F3C-4AEB-B757-87193EAD5207@fastcrypt.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-jdbc |
Additionally you should be aware, that using this mechanism once a
connection gets a hundred values, they are consumed even if you don't
use them.
Dave
On 7-Nov-05, at 4:18 PM, Kris Jurka wrote:
>
>
> On Mon, 7 Nov 2005, Mike Clements wrote:
>
>> I've got a bunch of PreparedStatements doing inserts on tables with
>> primary keys generated by sequences. For example, insert on table A,
>> take generated primary key, insert on table B assigning foreign key
>> generated value for A's primary key.
>>
>> For performance reasons, I need to batch these commands. But I
>> don't see
>> how it would be possible to continue using this approach with a
>> batch of
>> commands, because I need the results of the first insert to make the
>> second insert. Is there some way to do this or am I going to have to
>> stop using keygen and instead have my application generate its own
>> keys?
>>
>
> One option would be to tune your sequence generator to your batch
> size, consider: CREATE SEQUENCE myseq INCREMENT BY 100; Then if
> you fetch a nextval you know that you are also free to use the next
> 99 values as well in your batch statement without touching the
> sequence.
>
> Kris Jurka
>
> ---------------------------(end of
> broadcast)---------------------------
> TIP 6: explain analyze is your friend
>
From | Date | Subject | |
---|---|---|---|
Next Message | Andres Olarte | 2005-11-08 00:45:55 | Re: Batch with keygen? |
Previous Message | Kris Jurka | 2005-11-07 21:18:31 | Re: Batch with keygen? |