Re: [Geotools-gt2-users] GeoJSON - No such Attribute problem

[Justin D. <jde...@op...>]

On Fri, May 4, 2012 at 3:16 AM, Andrea Aime <and...@ge...>wrote:

> On Fri, May 4, 2012 at 2:18 AM, Gus MacAulay <gus...@gm...>wrote:
>
>> Hi All,
>>
>> I've been happily using the GeoJSON library but have run into a
>> problem.  Basically it doesn't seem to handle the case where features
>> have differing properties, for example;
>>
>> Reading this feature collection
>>
>> {"type":"FeatureCollection",
>>        "features":[
>>
>>  {"type":"Feature","geometry":{"type":"Point","coordinates":[144.5719,-37.6786]},"properties":{"A":"1"},"id":"points.1"},
>>
>>  {"type":"Feature","geometry":{"type":"Point","coordinates":[144.5976,-37.678]},"properties":{"B":"3"},"id":"points.3"}
>>        ]
>> }
>>
>> With something like this
>>
>> FeatureJSON fj = new FeatureJSON();
>> FeatureIterator<SimpleFeature> features =
>> fj.streamFeatureCollection(url.openConnection().getInputStream());
>>
>> while (features.hasNext()) {
>>            SimpleFeature feature = (SimpleFeature) features.next()
>> }
>>
>> Results in an exception
>>  java.lang.IllegalArgumentException: No such attribute:B
>>
>> Any thoughts? Is there a workaround?
>>
>
> As far as I know the parser assumes all the features have the same
> structure as the first one.
> Making it work otherwise is going to be hard since FeatureType are meant
> to be immutable...
> To deal with the above case the geotools geojson parser would have to be
> modified.
>
> One approach could be to parse the whole collection twice, the first time
> to determine
> the full attribute set and then to actually generate the features. This
> would be of course slow.
>
> Another approach could be to have a custom implementation that allows the
> set of attributes
> to be expanded along during the parsing, probably some custom subclass of
> SimpleFeatureType
> and maybe SimpleFeature too.
>

A couple of additional ideas.

One would be to accept a feature type definition before parsing and use
that for features. Obviously that would only work in cases where the
feature types and all possible attributes are known before hand.

Second would be to stick unrecognized attributes in the user data of the
feature rather than fail on parsing.

>
> Cheers
> Andrea
>
> --
> Ing. Andrea Aime
> GeoSolutions S.A.S.
> Tech lead
>
> Via Poggio alle Viti 1187
> 55054  Massarosa (LU)
> Italy
>
> phone: +39 0584 962313
> fax:      +39 0584 962313
> mob:    +39 339 8844549
>
> http://www.geo-solutions.it
> http://geo-solutions.blogspot.com/
> http://www.youtube.com/user/GeoSolutionsIT
> http://www.linkedin.com/in/andreaaime
> http://twitter.com/geowolf
>
>


-- 
Justin Deoliveira
OpenGeo - http://opengeo.org
Enterprise support for open source geospatial.