number of no NoData cells in raster

[martinmaretta]

Hi is there some way how to find out number of no NoData cells in raster? (without looping whole raster)  thanks a lot.

[RaphaelR]

there´s probably a better way to do this, but i´d try with numpy:

import arcpy
import numpy as np
import numpy.ma as ma

inRaster = r"E:\testraster.tif"

# convert inraster to numpy array, set nodata values to -9999
npArray = arcpy.RasterToNumPyArray(inRaster,"","","",-9999)
    
# mask the non-nodata values, and take them out by compressing
npM = ma.masked_where(npArray <> -9999, npArray).compressed()
    
# print how many cells are nodata    
print len(npM)

[RaphaelR]

maybe with numpy?

import arcpy
import numpy as np
import numpy.ma as ma


inRaster=r"E:\test1.tif"
# convert inraster to numpy array, set nodata values to -9999
npArray = arcpy.RasterToNumPyArray(inRaster,"","","",-9999)
    
# mask the non-nodata values, and take them out by compressing
npM = ma.masked_where(npArray <> -9999, npArray).compressed()
    
print npM.size
    

[AndrewChapkowski]

You might be able to use the bincount() function on numpy to do what you are looking for:
http://docs.scipy.org/doc/numpy/reference/generated/numpy.bincount.html
the values have to be non-negative though.

[DanPatterson_Retired]

Check out the IsNull tool in the spatial analyst toolset, it will produce a grid of 0's or 1's, simply open its table to get the cell count.

[curtvprice]

If it's an integer raster, you can calculate the number of data cells by summing up COUNT in the raster table and subracting it from rows * columns.

For floating grids, you must convert the cells to 0/1 with IsNull and sniff the COUNT value from the output raster RAT (I'd do that with a cursor for a table view generated for VALUE = 1).

[martinmaretta]

Thanks a lot, all of you