Hello,
I’m fairly inexperienced when it comes to GIS, so I wanted to find out how you all implement dynamic features in regards to Geoserver. I’m currently putting together a solution with MapBuilder and Geoserver (it seems like the best best choice), but I’m a little confused on how to handle features that will change often. MapBuilder doesn’t have very good support for SLD filters and test for on the fly GML (not from geoserver), which means which means I would have to rely on geoserver and a geodatabase in order to be able to get this kind of functionality. The problem is that I already have entities in my application that will be interfacing with this GIS solution and it seems silly to also store it in a geodatase. I also have feeds that will be converted to features on the map and they will change on a pretty frequent basis. So if I wrote those features to a geodatabase, it seems like I could end up losing a bit of performance there. How do you all handle non-geodatabases that already have some spatial data, and also how do you handle more dynamic data? Do you store this stuff in a geodatabase anyhow? If not, how do you render it based on attributes, such as with an SLD? Is there another option I am missing?
Thanks!
Ryan
Ryan Moquin ha scritto:
Hello,
I'm fairly inexperienced when it comes to GIS, so I wanted to find out how you all implement dynamic features in regards to Geoserver. I'm currently putting together a solution with MapBuilder and Geoserver (it seems like the best best choice), but I'm a little confused on how to handle features that will change often.
What do you mean by features that do change often? Is the contents
of the data source that changes, or its structure (that is, attributes
and the like)?
MapBuilder doesn't have very good support for SLD filters and test for on the fly GML (not from geoserver), which means which means I would have to rely on geoserver and a geodatabase in order to be able to get this kind of functionality. The problem is that I already have entities in my application that will be interfacing with this GIS solution and it seems silly to also store it in a geodatase. I also have feeds that will be converted to features on the map and they will change on a pretty frequent basis. So if I wrote those features to a geodatabase, it seems like I could end up losing a bit of performance there. How do you all handle non-geodatabases that already have some spatial data, and also how do you handle more dynamic data?
OGC services are build around the idea that the data structure is
not variable. You describe a feature type using an XSD schema, this
casts the structure of your data in stone. If your set of attributes
changes with time, you're going to basically break OGC clients
that are caching the feature type structure for performance reasons.
Anyways, let us know more about your use case and we'll see how to
handle it.
Cheers
Andrea
Ryan Moquin ha scritto:
I simply mean, let's say a status for a feed that I'm consuming that has geospatial data attached.
I think my question is how to start developing one of those schemas without a geodatabase to model after? I've been doing a lot of reading on various topics regarding GML and it's really confused me on exactly how to get started defining your own datastructure using it.
GeoRSS feeds are quite unstructured, and GML pretends structure. So yes, I guess you'd need to dump everything into a database.
This would be necessary as well because there is no datastore able
to read an RSS feed so you would miss the connection to your data source
as well.
Alternatively, you could develop a geotools datastore that reads the
feeds, provides structural information and access means for them,
and drop it into GeoServer (there is a plugin mechanism for datastore,
a properly coded one works out of the box).
Hope this helps, but I encourage anyone with more familiarity with
unstructured data to chime in.
Cheers
Andrea