Using bool array mask, replace False values with NaN

Question

I have two arrays,

a = array([
   [ 0.93825418,  0.60731973,  0.44218921,  0.90888805,  0.97695114],
   [ 0.27422807,  0.75870153,  0.12154102,  0.89137678,  0.04257262],
   [ 0.32855867,  0.17215507,  0.00302302,  0.95395069,  0.02596567],
   [ 0.18385244,  0.09108341,  0.27925367,  0.0177183 ,  0.41035188],
   [ 0.87229432,  0.73573982,  0.98554476,  0.72321398,  0.98316711],
   [ 0.16474265,  0.5308054 ,  0.27913615,  0.59107689,  0.6480463 ],
   [ 0.88356436,  0.22343885,  0.74900285,  0.43895017,  0.74993129],
   [ 0.08097611,  0.48984607,  0.33991052,  0.06431022,  0.10753135],
   [ 0.67351561,  0.13165046,  0.41327765,  0.21768539,  0.7337069 ],
   [ 0.65609999,  0.06241059,  0.3400624 ,  0.13234171,  0.23679716]
])

b = array([
   [False,  True,  True, False, False],
   [ True, False, False, False, False],
   [ True,  True, False, False, False],
   [False, False,  True, False,  True],
   [False, False, False,  True, False],
   [False,  True,  True,  True,  True],
   [False,  True, False,  True,  True],
   [False,  True,  True, False, False],
   [ True,  True,  True,  True,  True],
   [ True, False,  True, False,  True]
], dtype = bool)

Now I want using b to mask a, retain the True value in a, and replace the False value with NaN, getting a new array that has a shape like a.

How to do that?

As an alternative to what you specifically asked for, have you thought of doing something different? The package numpy.ma exists for this very reason. Note: in numpy.ma, a mask of True means the value is masked; you are using the opposite sense. — David Hammen
– David Hammen, Commented Aug 11, 2015 at 14:00

Alex Riley · Accepted Answer · 2015-08-11 13:55:49Z

12

You can use boolean indexing:

a[~b] = np.nan

This replaces all of values in a that correspond to False values in the mask b with np.nan:

>>> a
array([[        nan,  0.60731973,  0.44218921,         nan,         nan],
       [ 0.27422807,         nan,         nan,         nan,         nan],
       [ 0.32855867,  0.17215507,         nan,         nan,         nan],
       [        nan,         nan,  0.27925367,         nan,  0.41035188],
       [        nan,         nan,         nan,  0.72321398,         nan],
       [        nan,  0.5308054 ,  0.27913615,  0.59107689,  0.6480463 ],
       [        nan,  0.22343885,         nan,  0.43895017,  0.74993129],
       [        nan,  0.48984607,  0.33991052,         nan,         nan],
       [ 0.67351561,  0.13165046,  0.41327765,  0.21768539,  0.7337069 ],
       [ 0.65609999,         nan,  0.3400624 ,         nan,  0.23679716]])

edited Aug 11, 2015 at 13:55

answered Aug 11, 2015 at 13:46

Alex Riley

178k46 gold badges274 silver badges247 bronze badges

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

2 Comments

Holt Over a year ago

A small link that could be useful: docs.scipy.org/doc/numpy/reference/…

Alex Riley Over a year ago

Thanks @Holt - I've absorbed the link into the answer :-)

Collectives™ on Stack Overflow

Using bool array mask, replace False values with NaN

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related