Hi all We have the following scenario: every day we receive a meteorological dataset in NetCDF, that contains 104 subdatasets, each representing a different variable. Each of these subdataset has 10 bands, each representing the forecast value for that variable for the coming 10 days. We need to expose all of this datasets as different WMS and keep the access to all of the datasets (NetCDF) in the time.
The approach I was considering is the following: every day I would generate with a script a *different mapfile*, representing one single NetCDF dataset. Still with a script I would build 104 tiled shapefile, with a TIME field (WMS Time option). Each shapefile would contain 10 record, one for each band (day), set to the path of a virtual raster (still generated by the script) pointing to the correct combination of variable/day in the NetCDF file. The mapfile would contain 208 layers, two for each variable (the tile shapefile for time tiling the forecast day, and the raster for the variable). One alternative approach could be to create just a single WMS and append parameters in the request, and process them with *mapscript*, but this way I believe the service would go out from the OGC specification. Simply, I would like that any single day would have a specific WMS service, exposed by a different mapfile. This way the end user could query the specific WMS for a given day, using the TIME parameter for specifing the forecast day, and the LAYER parameter for specifing the variable. This way I will be able to build a request for each different dataset (represented by a different mapfile), each different variable (represented by a layer in mapfile) and each day forecast (represented by a single record in the tile shapefile). The request would be something like this: http://localhost/mapserver/global?MAP=/path/to/mapfiles/dataset_20111001.map&LAYERS=myvariable&TIME=2011-10-03&;... With Apache URL rewriting I could transform the request, making the reference to the mapfile not evident: http://localhost/mapserver/global/20111001?LAYERS=myvariable&TIME=2011-10-03&;... Do you think this could be a correct approach or there is a better way to proceed? Thanks P -- Paolo Corti Geospatial software developer web: http://www.paolocorti.net twitter: @capooti skype: capooti _______________________________________________ mapserver-users mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/mapserver-users
