Geographic Information Systems Stack Exchange is a question and answer site for cartographers, geographers and GIS professionals. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm using GDAL python to create a bunch of tiff files from a netcdf file.

def modelout_to_geotif(netcdfile, outdir, VarName):
    """ Given the path to the netcdf output,
    produces geotiffs for each band of the given variable name.

        netcdfile: string path to netcdf file
        VarName: string variable name in netcdf file (e.g. 'GridPrec')

        creates geotiffs for each 'band' (e.g. time step). Output is written
        to current working directory.

        nci = gdal.Open('NETCDF:{0}:{1}'.format(netcdfile, VarName))
        ncd = Dataset(netcdfile)
        print "Could not open input file: {0}".format(netcdfile)

        geotransform = nci.GetGeoTransform()
        print "Could not get geotransform.\n"
        projection = nci.GetProjection()
        print "Could not get projection.\n"

    numbands = nci.RasterCount
    times = [o2d(t) for t in ncd.variables['time']]
    print "Found {0} bands to process for {1}".format(numbands, VarName)
    for i in range(numbands):
        time = times[i]
        band = nci.GetRasterBand(i+1)
        raster = band.ReadAsArray()
        y,x = raster.shape

        output_file = "{0}_{1}.tif".format(VarName, time.strftime('%Y%m%d_%H'))
        subdir = os.path.join(outdir, VarName)
        if not os.path.exists(subdir):
        output_file = os.path.join(subdir, output_file)
        if os.path.exists(output_file):
        # Create gtif
        driver = gdal.GetDriverByName("GTiff")
        #TODO: Make this more generic
        dst_ds = driver.Create(output_file, x, y, 1, gdal.GDT_CFloat32 )

        # top left x, w-e pixel resolution, rotation, top left y, rotation, n-s pixel resolution
        dst_ds.SetGeoTransform( geotransform )

        # set the reference info 
        dst_ds.SetProjection( projection )

        # write the band

Now I would like to add those to a mosaic catalog in an ESRI File Geodatabase. I've managed to get it working using arcpy, but it's really slow, and I'd prefer to use GDAL. Is there a way? Below is my arcpy script:

import os
    import arcpy
    from arcpy import env
    raise ImportError("Unable to import arcpy - are you using ESRI Python?")

env = 'P:/Enki'

def add_to_mosaic(filegdb, raster_dir):    

    print env
    #Add Raster Dataset type Raster to FGDB Mosaic Dataset
    #Calculate Cell Size Ranges and Build Boundary
    #Build Overviews for Mosaic Dataset upon the 3rd level Raster Dataset pyramid
    #Apply TIFF file filter
    #Build Pyramids for the source datasets
    varnames = ['GridPrec']#,

    for VarName in varnames:
        gdbname = filegdb 
        mdname = os.path.join(gdbname, VarName)
        messages = 'Failed'
        sr =  arcpy.SpatialReference()    
# = 'WGS_1984_UTM_Zone_32N'
        sr.factoryCode = 32632
        print sr 
            noband = 1
            res = None
            #TODO: get from input data
            print 'trying to create %s' % VarName
            res = arcpy.CreateMosaicDataset_management(gdbname, mdname, sr, noband)

            print res.getMessages()

        rastype = "Raster Dataset"
        inpath = os.path.join(raster_dir, VarName)
        updatecs = "UPDATE_CELL_SIZES"
        updatebnd = "UPDATE_BOUNDARY"
        updateovr = "UPDATE_OVERVIEWS"
        maxlevel = "2"
        maxcs = "#"
        maxdim = "#"
        spatialref = "#"
        inputdatafilter = "*200701*.tif"
        subfolder = "NO_SUBFOLDERS"
        duplicate = "EXCLUDE_DUPLICATES"
        buildpy = "BUILD_PYRAMIDS"
        calcstats = "CALCULATE_STATISTICS"
        buildthumb = "NO_THUMBNAILS"
        comments = "Add Raster Datasets"
        forcesr = "#"

        res = arcpy.AddRastersToMosaicDataset_management(mdname,  rastype, inpath, updatecs, \
                                                   updatebnd, updateovr, maxlevel, maxcs, \
                                                   maxdim, spatialref, inputdatafilter, \
                                                   subfolder, duplicate, buildpy, calcstats, \
                                                   buildthumb, comments, forcesr)
        print res.getMessages()
share|improve this question
As far as I know, there is no raster support for the filegdb api released by esri. I wouldn't count on raster support any time, but who knows? – kyle May 29 '13 at 5:40
So you mean that the api released and accessible to gdal won't likely include the 'CreateMosaicDataset_management' functionality? – John May 29 '13 at 7:12
I think Kyle is right, the only way to manage Mosaic Datasets currently is through ArcPy. I've also run into performance bottleneck issues with this; it's quite frustrating since GDAL is so fast for reprojection etc. – greenlaw May 29 '13 at 14:19
I had started an ideas vote up thing for raster support, it garnered little support. My attempts to discuss it on arcmap forums was equally ignored. The FileGDB currently doesn't support rasters. – kyle May 31 '13 at 14:46

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.