Re: problem with new autocommit config parameter and jdbc
| От | Tom Lane |
|---|---|
| Тема | Re: problem with new autocommit config parameter and jdbc |
| Дата | |
| Msg-id | 15685.1031624823@sss.pgh.pa.us обсуждение исходный текст |
| Ответ на | Re: problem with new autocommit config parameter and jdbc (snpe <snpe@snpe.co.yu>) |
| Ответы |
Re: problem with new autocommit config parameter and jdbc
|
| Список | pgsql-hackers |
snpe <snpe@snpe.co.yu> writes:
> I'm use 'autocommit=false' and have problem with psql
> When any commnad is lost, then next commnad get error for transactions
> (simple select command).BTW
> snpe> select * from org_ba;
> ERROR: relation org_ba does not exists
> snpe> select * from org_ban;
> ERROR: current transactions is aborted, queries ignored until end of
> transaction block
Um, what's wrong with that?
It seems to me that an application that is using autocommit=off will
expect the first SELECT to start a transaction block. If the first
SELECT fails, then subsequent commands *should* fail until you commit
or rollback. Certainly if you did an explicit BEGIN before the first
SELECT, the above is what you'd get --- why should implicit BEGIN
work differently?
regards, tom lane
В списке pgsql-hackers по дате отправления: