creating a boolean array which compares numpy elements to None

Question

I have a numpy array with dtype=object, and I want to create a boolean array identifying which elements are None. But it looks like None behaves differently...

a = np.array(['Duck','Duck','Duck','Goose',None,1,2,3,1,3,None,4])
print a == 'Duck'
print a == 3
print a == None

which results in

[ True  True  True False False False False False False False False False]
[False False False False False False False  True False  True False False]
False

Is there an "numpythonic" way to get a boolean array of the None elements? I can use

np.array([x is None for x in a])

but this seems like there should be a better way.

I definitely would not have expected ndarrays to behave differently for None... good catch! — SethMMorton
– SethMMorton, Commented Sep 25, 2014 at 22:38
...especially because of this quote from the NumPy docs: "Each of the arithmetic operations (+, -, *, /, //, %, divmod(), ** or pow(), <<, >>, &, ^, |, ~) and the comparisons (==, <, >, <=, >=, !=) is equivalent to the corresponding universal function (or ufunc for short) in Numpy." I can only assume the np.equal is the ufunc for ==. — SethMMorton
– SethMMorton, Commented Sep 25, 2014 at 22:44

Ashwini Chaudhary · Accepted Answer · 2014-09-25 22:36:55Z

13

You can use numpy.equal:

In [20]: np.equal(a, None)
Out[20]: 
array([False, False, False, False,  True, False, False, False, False,
       False,  True, False], dtype=bool)

answered Sep 25, 2014 at 22:36

Ashwini Chaudhary

252k60 gold badges478 silver badges519 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

SethMMorton Over a year ago

This is definately a good workaround, but do you know why it doesn't work for None? The numpy docs say that == should match it's ufunc equivalent, and I imagine the equal is the matching ufunc.

Ashwini Chaudhary Over a year ago

@SethMMorton Found a related ticket on Github: == None seems to do is None for the array instead of returning element-wise comparison #3759

SethMMorton Over a year ago

Awesome... I guess that solves the why (it's a bug). Thanks!

sebix Over a year ago

Comparison with == None should give a FutureWarning: comparison to 'None' will result in an elementwise object comparison in the future. in current version. None is the only exception where this happens, but that will get fixed.

Collectives™ on Stack Overflow

creating a boolean array which compares numpy elements to None

1 Answer 1

4 Comments

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Linked

Related