Re: [JDBC] JDBC behaviour - Mailing list pgsql-general

From Sridhar N Bamandlapally
Subject Re: [JDBC] JDBC behaviour
Date
Msg-id CAGuFTBV14KqA_DmX-_=xefZq7QKn0DER73HbX+dkDehrssqtuw@mail.gmail.com
Whole thread Raw
In response to Re: [JDBC] JDBC behaviour  (Vladimir Sitnikov <sitnikov.vladimir@gmail.com>)
Responses Re: [JDBC] JDBC behaviour
Re: [JDBC] JDBC behaviour
List pgsql-general
If we want transactions in "begin-end" then its fine, 

but in this case all these transactions are independent with autocommit off,

user choice to continue with commit or rollback

Thanks
Sridhar



On Thu, Feb 18, 2016 at 1:43 PM, Vladimir Sitnikov <sitnikov.vladimir@gmail.com> wrote:
>Is it a bug or do we have other any alternate way to handle this ?

PostgreSQL is strongly against "partial commits to the database". If
you think a bit about it, it is not that bad.
You got an error, what is the business case to commit the partial
transaction then?

Exceptions should not be used for a "control flow", should they?


If you want to shoot yourself in a foot for fun and profit, you can
try https://github.com/pgjdbc/pgjdbc/pull/477.
What it does, it creates savepoints before each statement, then it
rollbacks to that savepoint in case of failure.

Vladimir

pgsql-general by date:

Previous
From: Victor Blomqvist
Date:
Subject: Query plan not updated after dropped index
Next
From: Vladimir Sitnikov
Date:
Subject: Re: [JDBC] JDBC behaviour