Jesper Krogh <jesper(at)krogh(dot)cc> wrote:
> Is there any way I can force explicit use of transactions on
> I do have users sitting on the system, that may forget to code it
> that way and it would be nice to be able to force the database to
> just kick them out if it didn't happen.
Not that I can think of, unless you pass the queries through some
proxy which enforces the rule. Generally, a statement outside of an
explicit transaction block starts its own transaction for just that
statement, just as if the statement were wrapped in BEGIN/COMMIT.
You could maybe get glimpses of such statements by periodically
looking at pg_stat_activity for statements which have the same
timestamp for xact_start and query_start.
In response to
pgsql-admin by date
|Next:||From: Nicholson, Brad (Toronto, ON, CA)||Date: 2012-01-23 13:10:26|
|Subject: Re: Best practise for upgrade of 24GB+ database|
|Previous:||From: c k||Date: 2012-01-21 18:07:40|
|Subject: Re: grouping of query data in xml|