Python Get minimum value without -inf

Question

I am trying to find the minimum of an array without -inf values. What I could find there is numpy function to filter out NaNs but not -inf:

import numpy as np
np.nanmin([1, 2, np.nan, np.NINF])

The desired output is:

1

Answer 1

Select the non -Infs and get the min ignoring NaNs with np.nanmin -

np.nanmin(a[a != -np.inf])

Sample run -

In [209]: a
Out[209]: array([  1.,   2.,  nan, -inf])

In [210]: np.nanmin(a[a != -np.inf])
Out[210]: 1.0

Answer 2

Use np.isfinite to get only real values:

import numpy as np
a = np.array([1, 2, np.nan, np.NINF])
min_value = a[np.isfinite(a)].min()
# >>> 1.0

Answer 3

An alternative approach using masked arrays:

>>> import numpy as np
>>> np.ma.masked_invalid([1, 2, np.nan, np.NINF]).min()
1.0

masked_invalid will ignore Inf and NaN values.