On 09.12.2022 09:46, Pavel Stehule wrote:
> I don't like the idea about possible replacement of INTO STRICT by INTO
> + extra warnings.
>
> Handling exceptions is significantly more expensive than in Oracle, and
> using INTO without STRICT with the next test IF NOT FOUND THEN can save
> one safepoint and one handling an exception. It should be mentioned in
> the documentation. Using this very common Oracle's pattern can have a
> very negative impact on performance in Postgres. If somebody does port
> from Oracle, and wants compatible behavior then he should use INTO
> STRICT. I think it is counterproductive to hide syntax differences when
> there is a significant difference in performance (and will be).
Fair enough. Thank you, Pavel.
--
Sergey Shinderuk https://postgrespro.com/