A mole from the geoprocessing team sent me this neat example.This method could be used to generate $$XMAP and $$YMAP using the environment raster extent and cell size.Sure would be nice to include a script tool in the Arc 10 toolbox to both serve as a simple example of how to use NumpyArrayToRaster, and keep the functionality in the standard product. Otherwise the handfull of geeks that use this functionality will have to do it ourselves. Here's the (now trivial, with the example below) tool I will probably write if you don't release one:CreateLocationRaster <Out_raster> {ROWMAP | COLMAP | XMAP | YMAP} {Template_raster}
## NumPyArrayToRaster (in_array, {lower_left_corner},
## {x_cell_size}, {y_cell_size}, {value_to_nodata})
import os, sys, arcpy
import numpy as NUM
# Set path
sPath = sys.path[0]
outPath = os.path.join(os.path.dirname(sPath),"output")
# Set environment
arcpy.env.workspace = outPath
arcpy.env.overwriteOutput = 1
#------------------------------------------
# Create and populate an array
myArray = NUM.zeros((10,16),dtype='int16')
# Fill row with row number
for i in range(0,10):
myArray = i + 1
print myArray
# Create a raster
rowRas = arcpy.NumPyArrayToRaster(myArray)
rowRas.save("rowRas")
print "...saved rowras"
#------------------------------------------
# Fill col with col number
for j in range(0,16):
myArray[:,j] = j+1
print myArray
# Create a raster
colRas = arcpy.NumPyArrayToRaster(myArray)
colRas.save("colRas")
print "...saved colras"
And here's another (certifiably cool) way to do it, sent by a USGS colleague... using the NumPy "indices" method:
import numpy as np
import arcpy
# Setup some rasters of rows and columns
# (like old $$ROWMAP and $$COLMAP)
nprows = np.indices((10,10))[0]
npcols = np.indices((10,10))[1]
# Convert the numpy arrays to ESRI rasters (ie Raster objects)
# called 'rows' and 'cols'
rows = arcpy.NumPyArrayToRaster(nprows)
cols = arcpy.NumPyArrayToRaster(npcols)
