Thanks Frank for these remarks. I work on pds driver (in fact the subdriver ISIS2). I agree with you that "some folks would want to distinguish the various kinds of nodata values".
And it's the reason i would like the default behaviour is "we have various no data value" but we can add a flag option for that we try to open a file we can merge all no data value. I had read the rfc15 and i will think about this Best Regards Ludovic On 8 déc, 18:46, Frank Warmerdam <warmer...@pobox.com> wrote: > I have lost track of what driver we are discussing. It is certainly > possible for a theory to remap various nodata values to a single one > if that is perceived as appropriate for that driver. However, in the > situation you described it seems plausible that some folks would want > to distinguish the various kinds of nodata values and that merging > them is interfering with access to the original raster data in a way > that might be detrimental. > > I hesitate to bring it up, but another option is to return a validity > mask for the dataset derived by checking for the various nodata values. > There is already automated support for generating validity masks from > regular bands with single nodata values built into the core. One > downside to this approach (not counting complexity) is that relatively > few applications honour the mask concept at this time. > > http://trac.osgeo.org/gdal/wiki/rfc15_nodatabitmask > > Best regards, > -- > ---------------------------------------+-------------------------------------- > I set the clouds in motion - turn up | Frank Warmerdam, warmer...@pobox.com > light and sound - activate the windows |http://pobox.com/~warmerdam > and watch the world go round - Rush | Geospatial Programmer for Rent > > _______________________________________________ > gdal-dev mailing list > gdal-...@lists.osgeo.orghttp://lists.osgeo.org/mailman/listinfo/gdal-dev _______________________________________________ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev