Numpy array mean function did not exclude masked element in mean computation

Question

I am new in python programming, please forgive me if my question is too basic. But I am trying to use masked_array to calculate the mean of three arrays to produce a third array without using the elements with values less than zero. Using these steps:

import numpy as np
from numpy.ma import masked_array

d=[]
a = np.array([[-2,-3,-4,-6],[5,2,6,1],[9,3,2,4],[3,1,1,2]])
b = np.array([[3,4,2,4],[5,2,6,1],[9,3,2,4],[0.3,12,1,3]])
c = np.array([[2,3,4,5],[7,0,1,5],[10,9,2,3],[1.5,2.01,2,0.2]])

mask = (a <= 0).astype(int) 
a = masked_array(a,mask)

d.append(a)
d.append(b)
d.append(c)
result = np.array(d).mean(axis=0)
print result

[[ 1.          1.33333333  0.66666667  1.        ]
[ 5.66666667  1.33333333  4.33333333  2.33333333]
[ 9.33333333  5.          2.          3.66666667]
[ 1.6         5.00333333  1.33333333  1.73333333]]

I thought that (from documentation that) if one uses the numpy array mean function on arrays with masked values, that it will not consider them (the masked values) in computing the mean. I expected the result to be

[[  2.5          3.5  3.  4.5        ]
 [ 5.66666667  1.33333333  4.33333333  2.33333333]
 [ 9.33333333  5.          2.          3.66666667]
 [ 1.6         5.00333333  1.33333333  1.73333333]]

Please, does anyone have some tips on how I can use the numpy.ma.masked_array to archive this?

This isn't part of the problem, but you could write d = [a,b,c] instead of appending. — hpaulj
– hpaulj, Commented Jul 31, 2015 at 15:00

Warren Weckesser · Accepted Answer · 2015-07-31 14:23:38Z

1

The problem is that np.array(d) does not create a masked array. The mask in a is lost when the arrays in the list d are assembled into a bigger (nonmasked) array.

One way to fix it is to replace this:

result = np.array(d).mean(axis=0)

with

result = masked_array(d).mean(axis=0)

E.g.:

In [27]: result = masked_array(d).mean(axis=0)

In [28]: result
Out[28]: 
masked_array(data =
 [[2.5 3.5 3.0 4.5]
 [5.666666666666667 1.3333333333333333 4.333333333333333 2.3333333333333335]
 [9.333333333333334 5.0 2.0 3.6666666666666665]
 [1.5999999999999999 5.003333333333333 1.3333333333333333
  1.7333333333333334]],
             mask =
 [[False False False False]
 [False False False False]
 [False False False False]
 [False False False False]],
       fill_value = 1e+20)

answered Jul 31, 2015 at 14:23

Warren Weckesser

116k20 gold badges207 silver badges224 bronze badges

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

1 Comment

Kela Over a year ago

Thank your very much @Warren Weckesser. This work perfectly for this scenario.

Collectives™ on Stack Overflow

Numpy array mean function did not exclude masked element in mean computation

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related