[mapserver-dev] Re: WFS GetFeature Performance - BBOX vs. Filter

[Brad Posthumus]

Yewondwossen Assefa wrote:
> 
> For your particular case, if the filter encoding is "simple" (simple in 
> this case means one bbox, (one spatial expression), that limits the 
> whole query), you can certainly set the query rectangle to this bbox.  
> You can enter a bug and patch on it and assign it to me (assefa) and I 
> will comment/review/apply it.

Thanks for taking the time to respond and for offering to look into this
further. I logged the bug with a suggested patch
(http://trac.osgeo.org/mapserver/ticket/4049). 

The patch tries to account for the different scenarios where using the query
extent wouldn't be valid (e.g. table selections) but I can't say for certain
if it addresses for all of them. It also doesn't account for all scenarios
where a smaller extent would work (e.g. table selection AND query extent).
So for many cases, it simply uses the map extent as before. But when it does
use the query extent, it makes a huge difference in performance.

I tried to make it as unobtrusive as possible by only inserting code into
the one function without changing anything that's existing. C is definitely
not my strong point so there's likely better ways to code this; feel free to
chainsaw it to pieces if it helps!

Thanks again!

--
View this message in context: http://osgeo-org.1803224.n2.nabble.com/WFS-GetFeature-Performance-BBOX-vs-Filter-tp6890106p6902280.html
Sent from the Mapserver - Dev mailing list archive at Nabble.com.