[pycsw-devel] repository geometry storage

Angelos Tzotsos gcpp.kalxas at gmail.com
Wed Apr 6 17:19:05 EDT 2011


Hi Tom,

To really answer this question, we have to make a benchmark between 
bbox, wkt and wkb.

But my feeling is that when we create the PostGIS modules, we will need 
to use the native PostGIS geometries for performance. In the same 
manner, using a native format for Shapely, theoretically would cut down 
transformation time and keep only processing time during spatial 
queries. But is the WKT the *real* native format for Shapely? Or perhaps 
WKB? Or perhaps something internal? We have to look into this.

I think for the time, we should stick with the current storage model for 
0.1 release, and focus on this for later releases.

Best regards,
Angelos

On 04/06/2011 10:01 PM, Tom Kralidis wrote:
>
> Hi: FYI in trunk, we have the requirement for storing geometries in 
> the format 'minx,miny,maxx,maxy'.
>
> This format is adhered to for:
>
> - importing metadata into the repository
> - searching metadata spatially (server.util.query_bbox() and 
> server.util.bbox_not_query())
> - presenting metadata
>
> For spatial searching, we are using Shapely [1] to perform operations 
> like intersects (ogc:BBOX) and disjoint (ogc:Not/ogc:BBOX).  We 
> transform the geometry to wkt (server.util.bbox2wkt) before passing to 
> Shapely.
>
> Is there benefit to storing geometries as WKT in the repository?  Or 
> WKB (maybe for better performance)?  Or something else?
>
> The change would be easy to implement, the question is would this be 
> of value, in the context of the life cycle of catalog metadata 
> (import/search/present).
>
> Comments?
>
> ..Tom
>
> [1] http://trac.gispython.org/lab/wiki/Shapely
>
>
> ------------------------------------------------------------------------------
> Xperia(TM) PLAY
> It's a major breakthrough. An authentic gaming
> smartphone on the nation's most reliable network.
> And it wants your games.
> http://p.sf.net/sfu/verizon-sfdev
>
>
> _______________________________________________
> pycsw-devel mailing list
> pycsw-devel at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/pycsw-devel


-- 
Angelos Tzotsos
Remote Sensing Laboratory
National Technical University of Athens
http://users.ntua.gr/tzotsos

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/pycsw-devel/attachments/20110407/e95fa03a/attachment.html


More information about the Pycsw-devel mailing list