Hello,
I am trying to do a few Spatial Analyst tasks using annual composites of night lights around the globe. These are ~2.8GB tif files. One such task that seems imposible thus far is Extracting a Sample. My Python Script is a follows:
>>> import arcpy
>>> from arcpy.sa import *
>>> arcpy.env.workspace = r'C:\Users\Admin.LoveLeePC\Documents\ArcGIS\Default.gdb'
>>> Sample('ChngTmsLSPop1.tif','sampleLocation',r'C:\Users\Admin.LoveLeePC\Documents\ArcGIS\FinalNoLightStats\sample',"NEAREST")
This particular run gave the error:
Runtime error <class 'arcgisscripting.ExecuteError'>: ERROR 010070: Raster read error. MSEEK: No error ERROR 010067: Error in executing grid expression. Failed to execute (Sample).
This error came after the program maxed out one of the cores on my core 2 quad CPU for about 5 hours... The second I clicked the GUI of the program, it through this error. It seems adding any threads just shuts down the app in a second...
I know Arc is single threaded, but is there some way (maybe multiprocessing controls in Python) to take advantage of the rest of my CPU?
Am I crazy to think that the size of this raster is the issue? Is this how the Sample tool is meant to work?
Even when I sample much smaller areas, my computer crashes pretty quick. I have similar experiences with map algebra, and especially ZonalStatisticsAsTable.
Finally, I am hesitant to believe what I am saying since I do have a quad core with 8 GB of DR3 RAM, a very nice media-PC graphics card, etc.
Can someone please explain this to me???
Thanks,
Ephy
