[postgis-devel] Removing geometry parsing checks for data loading

[Mark Cave-Ayland]

David Fuhry wrote:

>    I agree that the shapefiles with unclosed polygons are totally 
> invalid.  But I need those records' attributes, and  their siblings 
> records' valid geometries in the db.  Besides shapefile preprocessing or 
> loading all geom records as "bytea"s, relaxing the geometry parser is 
> the least-hackish way I see to bulk load them.

At the end of the day, the input data is invalid and so if I were in 
your position then I would either a) fix the broken application which is 
producing these invalid shapefiles or b) complain to the people who are 
supplying the invalid data. I don't particularly see this as a PostGIS 
issue I'm afraid :(

> Similar ambiguities exist for the "num_points % num_dimensions == 0" and 
> "num_poly_vertices >= 3" checks.  And shp2pgsql's logic would always 
> have to be synced with the (E)WKB parser's.  (To me) seems to be a 
> slippery slope to add geometry validation/modification to shp2pgsql.

Well, the recent work in SVN HEAD means that the parser is now built as 
a separate library. The theory is that other utilities including 
shp2pgsql will be able to make use of this to avoid code 
duplication/improve consistency. So this won't be an issue moving forward ;)


ATB,

Mark.

-- 
Mark Cave-Ayland
Sirius Corporation - The Open Source Experts
http://www.siriusit.co.uk
T: +44 870 608 0063