[postgis-users] inlined SQL functions

[Paul Ramsey]

Certainly SQL has been available as a language for a very very long time 
(I wrote SQL procedures in version 6.4).  The only problem with making 
the index operation inline for the OGC functions is that we are being 
"smart" for the user, and may end up creating ugly SQL that confuses the 
planner and makes things worse in some cases.

P

Markus Schaber wrote:
> Hi, Strk,
> 
> strk at refractions.net wrote:
> 
> 
>>>This is interesting, thanks for pointing out!
>>>I'll make some tests on older postgresql versions.
>>
>>It seems this feature is only available starting
>>at PostgreSQL 7.4 (found in release notes).
> 
> 
> Does that mean that older PostgreSQL versions don't provide the "SQL"
> procecural language, or that they don't inline them?
> 
> We could write small SQL wrappers around "touches", "intersects",
> "contains" etc. that prepend the actual function call with a && check to
> allow the query planner to do its work. This should pay off for most but
> academic cases.
> 
> Markus
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users