arcpy cursor not accepting float(Inf)

Question

I have a vector layer in which I have added some fields of type 'float' using arcpy.AddField_management. I'm performing simple calculations between those fields using update cursors but the problem is that when I divide the values by 0, I get a run-time error even though I'm using Inf with the float function.

This is my code:

with arcpy.da.UpdateCursor('flurstcuke_cropped5', ['FID','totresgeb', 'totresgbpt','difresarea','scaledif']) as cursor6:

for row6 in cursor6:

   row6[3] = row6[2] - row6[1]

   try: 

       row6[4] = ((row6[2] - row6[1])/row6[1]) * 100

   except ZeroDivisionError:

      row6[4] = float('INF')


   cursor6.updateRow([row6[0], row6[1], row6[2], row6[3], row6[4]])

The message I get is that 'the value type is incompatible with the field type. [scaledif].' And I know it's referring to Inf since I have also set up a counter in my script to determine where the program is. But I don't understand that since 'Inf' is a floating number here and the field's type is also float

Answer 1

float("Inf") is not a true float. You can't do normal math with it. Feature class fields can only store actual numeric values, within specific ranges. From the help:

Data type     |  Storable range
---------------------------------------------------
Short integer |  -32,768 to 32,767
Long integer  |  -2,147,483,648 to 2,147,483,647
Float         |  Approximately -3.4E38 to 1.2E38
Double        |  Approximately -2.2E308 to 1.8E308

If you want the largest possible value, you should be able to import sys and use row6[4] = sys.float_info.max (although this is the size of the largest double in ArcGIS, not the largest float; the distinction between floats and doubles is blurred in Python).