Trying to get my head around this...
a = float("NaN")
a == float("NaN")
Out: False
np.isnan(a)
Out: True
Asked
Active
Viewed 64 times
0
-
8Because `NaN` is not equal to anything, including itself. By definition. – tkausl Jul 17 '22 at 15:32
-
NaN is effectively a null value: https://wiki.dataself.com/NULL_Values – Barry Carter Jul 17 '22 at 15:34
-
1Related post - https://stackoverflow.com/questions/944700/ – Daniel Hao Jul 17 '22 at 15:36
-
There are really many different not-a-number values, `float("NanN")` just produces one of them. One of the more significant consequences of using inband signaling of errors is that floating-point equality is *not* an [equivalence relation](https://en.wikipedia.org/wiki/Equivalence_relation), because reflexivity is not respected. – chepner Jul 17 '22 at 15:41
1 Answers
0
"NaN" means that it isn't equal to anything. So {thing1 which is NaN} == {thing2 which is NaN} will always be "False" because NaN is not equal to anything, including itself.
Daisy Welham
- 225
- 1
- 10
-
There is no point to *repeat* other comments here. Even better is to include some more related reference..to make it complete. – Daniel Hao Jul 17 '22 at 15:40
-
1