topic Re: Not a number (nan/1.#QNAN) condition testing for a FGDB raster attribute table? in Python Questions

Not a number (nan/1.#QNAN) condition testing for a FGDB raster attribute table?

ChrisSnyder — Mon, 17 Sep 2012 21:46:30 GMT

I'm cursoring through a FGDB raster attribute table in v10.1... Some of the attribute values are "Null" (aka 1.#QNAN), but when when evaluated in a cursor, they appear as a value called nan, for example:

>>> print updateRow[updateRows.fields.index("HCB")] nan

Unfortunatly nan is not a string but a float value... which is what the HCB field type is.

>>> type(updateRow[updateRows.fields.index("HCB")]) <type 'float'>

And nan isn't the same as None

>>> updateRow[updateRows.fields.index("HCB")] == None False

Problem is, I need a way to test for these nan values and then evaluate them as 0's... Something like:

hcbValue = updateRow[updateRows.fields.index("HCB")]) if hcbValue == None:     hcbValue = 0

If these evaluated as None, then of course it would be easy... but they don't, so it aint.

Anyone know how to do a conditional test for nan values???

Re: Not a number (nan/1.#QNAN) condition testing for a FGDB raster attribute table?

ChrisSnyder — Mon, 17 Sep 2012 21:57:35 GMT

Found it

math.isnan()

as in

>>> math.isnan(updateRow[updateRows.fields.index("HCB")]) True

Re: Not a number (nan/1.#QNAN) condition testing for a FGDB raster attribute table?

ChrisSnyder — Sat, 11 Dec 2021 20:11:22 GMT

FYI: A ESRI nan is not the same as

float('nan')
or
str(1e400*0)

...which I guess are other/older methods of evaluating nan values (http://stackoverflow.com/questions/944700/how-to-check-for-nan-in-python)

So...

>>> updateRow[updateRows.fields.index("HCB")] in [float('nan'), str(1e400*0)]
False

but the math.isnan() works...