CODEWITHSUNDEEP
pythonnumpy
user3362595
user3362595

Reputation: 31

How to mask an array column by column

I have an array like the following

Table = [[True,  False, False, False, False], 
         [False, False, True,  False, False],
         [True,  False, False, True,  False]]

Is there any way to mask all column by column to MaskTable, the final result as 

MaskTable = [True, False, True, True, False]

Upvotes: 1

Views: 84

Answers (2)

Alain T.
Alain T.

Reputation: 42143

If you are not using numpy, you can filter the columns using zip() within a list comprehension:

Filtered = [ [v for v,m in zip(row, MaskTable) if m] for row in Table ]

If you do use numpy, your table and mask must be numpy objects:

import numpy as np

Table = np.array([[True,  False, False, False, False], 
         [False, False, True,  False, False],
         [True,  False, False, True,  False]])
MaskTable = np.array([True, False, True, True, False])

Filtered = Table[:,MaskTable]

The above is the answer to your question as formulated in the title.

Your sample output, however, suggest that what you are actually trying to do is to apply an OR relation column-wise in your matrix:

MaskTable = np.any(Table,axis=0)

or

MakTable  = np.logical_or.reduce(Table,axis=0)

Upvotes: 0

ClearStream
ClearStream

Reputation: 31

Try:

MaskTable = np.any(Table, axis = 0)

Output:

array([ True, False,  True,  True, False])

This uses numpy.any(), which checks if any of the values along the given axis evaluate to True. numpy is imported as np here.

Upvotes: 2

Related Questions