Arcpy Bug? loop error ..runs great for 899 iterations and then fails at the 900th

[EricScharnhorst]

Before it breaks, my script does this in a loop:

for i in range (1, 2000):

if ID = i:

step 1: export one point whose ID = i, feature class to feature class
step 2: make polygon, 500 meter buffer around the point
step 3: clip a raster with the polygon from step 2
step 4: do a bunch of spatial analysis on the clipped raster
step 5: export a table with stats from step 4

The loop breaks at step 2 after the 900th iteration, which is just about 1/2 way to completion.

Here is the error message:
'File "C:\Programmer\ArcGIS\Desktop10.0\arcpy\arcpy\management.py", line 7692, in Clip
   raise e
arcgisscripting.ExecuteError: ERROR 000622: Failed to execute (Clip). Parameters are not valid.
ERROR 000628: Cannot set input into parameter rectangle.

The errors makes sense only because step 2 does not produce a polygon after the 900th time through the loop. Question is, why no polygon?

Thanks,
Eric

[ChrisMathers]

If the buffer isnt making a poly, what are the values for that record that are being put into the buffer?

[ChrisSnyder]

What is your memory consumption upon the 899th loop? Perhaps you are running out of memory.

[EricScharnhorst]

If the buffer isnt making a poly, what are the values for that record that are being put into the buffer?

Values are good ..that record, and those that follow it all work when using the buffer tool by hand in ArcMap.

[EricScharnhorst]

What is your memory consumption upon the 899th loop? Perhaps you are running out of memory.

Memory is good. The loop breaks at 900 even if I start at 899, for example .."for i in range (899, 2000):"

[ChrisMathers]

Does it fail if you start at 900? How about 901? I run cursors on 117000 record parcel feature classes every Monday morning with no problem. Can we see the whole of the code?

[EricScharnhorst]

Does it fail if you start at 900? How about 901? I run cursors on 117000 record parcel feature classes every Monday morning with no problem. Can we see the whole of the code?

Yes, it fails also when starting at 900 or after 900.

[ChrisMathers · Accepted solution]

That sounds like an error in your source point file then. Have you tried Repair Geometry on the point file?

Also, if you are using a cursor to iterate over a point file and exporting the points one at a time to a new feature class just for the geometry you can chop that step out by using the geometry object. The row object can give your the geometry to use as the input for the buffer tool. This would save the time it takes to write out to a new file over and over again.

[EricScharnhorst]

That sounds like an error in your source point file then. Have you tried Repair Geometry on the point file?

Also, if you are using a cursor to iterate over a point file and exporting the points one at a time to a new feature class just for the geometry you can chop that step out by using the geometry object. The row object can give your the geometry to use as the input for the buffer tool. This would save the time it takes to write out to a new file over and over again.

Bad news is Repair Geometry didn't change anything

Good news is the geometry object is working as advertised. Thanks for the tip, Chris. The loop is continuing now past 900, and each record is processed faster ..by about 2 seconds. This move will save the loop 1 hour in total.

[ChrisMathers]

Your welcome. Happy to help.