Re: BUG #11671: UNACCENT call with constant arg evaluated for each row

Tom Lane-2 wrote
> t.chaumeny@

>  writes:
>> Since UNACCENT function
>
> [...] than naive SQL [...]

Was that intentional... ;)

Given the planner knows its safe, and has to do a sequential scan and so
evaluate the expression many times, the question becomes why then doesn't it
choose that which is, in reality, the faster plan?

Reading 9.3@35.6
http://www.postgresql.org/docs/9.3/static/xfunc-volatility.html

I take it had the plan used an index instead of a sequential scan the
optimization would have occurred...

David J.




--
View this message in context:
http://postgresql.1045698.n5.nabble.com/BUG-11671-UNACCENT-call-with-constant-arg-evaluated-for-each-row-tp5823010p5823070.html
Sent from the PostgreSQL - bugs mailing list archive at Nabble.com.