On 2/20/2016 2:44 AM, Craig Ringer wrote:
> Hand-waving about how we're doing it wrong won't get you anywhere.
>
near as I can tell, the OP has used some sort of SQL (unspecified) where
multiple inserts within a transaction are individually inserted,
regardless of one failing.
to me this seems to break the rules of transaction semantics, but it
matches the behavior with autocommit=on ...
--
john r pierce, recycling bits in santa cruz