| From: | Kris Jurka <books(at)ejurka(dot)com> |
|---|---|
| To: | Dave Cramer <pg(at)fastcrypt(dot)com> |
| Cc: | Oliver Jowett <oliver(at)opencloud(dot)com>, Paul Tomblin <ptomblin(at)gmail(dot)com>, pgsql-jdbc(at)postgresql(dot)org |
| Subject: | Re: Atomic operations? |
| Date: | 2008-03-20 03:43:26 |
| Message-ID: | Pine.BSO.4.64.0803192335220.29669@leary.csoft.net |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-jdbc |
On Wed, 19 Mar 2008, Dave Cramer wrote:
>
> On 18-Mar-08, at 7:18 PM, Oliver Jowett wrote:
>
>> Actually it's fine the implicit transaction created by the server to
>> support autocommit wraps both queries in a single transaction and does not
>> commit/rollback until the end of the second query.
>>
> Yeah, using v3 protocol this is presented as one prepare/execute. Thanks for
> catching that.
For the record, v2 sends both statements in a single query message
grouping them in the same transaction. v3 sends a prepare and execute for
each statement (as required by the extended query protocol), but the
transaction is demarcated by the Sync message and there's only one of
those.
Kris Jurka
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Craig Ringer | 2008-03-24 05:45:05 | Non-ORM layers over JDBC |
| Previous Message | Kris Jurka | 2008-03-20 03:35:05 | Re: JDBC rewriting a bad query? |