Extracting specific columns in numpy array by condition

Question

I have a homework assignment to extract a 2-dimensional numpy array out of another 2-dimensional np array by choosing specific columns by condition (not by range).

So I have an array A with shape (3, 50000). I am trying to get a new array with shape (3, x) for some x < 50000 with the original columns ofAthat satisfy the third cell in the column is-0.4 < z < 0.1`.

For example if:

A = [[1,2,3],[2,0.5,0],[9,-2,-0.2],[0,0,0.5]]

I wish to have back:

B = [[2,0.5,0],[9,-2,-0.2]

I have tried to make a bool 1 rank array that holds true on the columns I want, and to some how combine between the two. The problem it's output is 1 rank array which is not what I am looking for. And I got some ValueErrors..

bool_idx = (-0.4 < x_y_z[2] < 0.1)

This code made some troubles:

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

I can do it with some loops but NumPy got so many beautiful function I am sure I am missing something here..

What I try my methos it returns a 1-rank array.. I tried to use x_y_z[bool_idx] when bool_idx is 1-rank array with true on the right columns and false on the ones I dont want.. but it won't work like that.. — Shaq
– Shaq, Commented Apr 4, 2019 at 7:26
What do you mean by full runnable example? What I showed with A and B is not clear enough? — Shaq
– Shaq, Commented Apr 4, 2019 at 7:59

Mad Physicist · Accepted Answer · 2019-04-04 08:04:41Z

3

In Python, the expression -0.4 < x_y_z[2] < 0.1 is roughly equivalent to -0.4 < x_y_z[2] and x_y_z[2] < 0.1. The and operator decides the truth value of each part of the expression by converting it into a bool. Unlike Python lists and tuples, numpy arrays do not support the conversion.

The correct way to specify the condition is with bitwise & (which is unambiguous and non-short-circuiting), rather than the implicit and (which short circuits and is ambiguous in this case):

condition = ((x_y_z[2, :] > - 0.4) & (x_y_z[2, :] < 0.1))

condition is a boolean mask that selects the columns you want. You can select the rows with a simple slice:

selection = x_y_z[:, condition]

answered Apr 4, 2019 at 8:04

Mad Physicist

116k29 gold badges202 silver badges292 bronze badges

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

1 Comment

Shaq Over a year ago

Yesit works! Thank you so much! I dont understand tho what is the meaning of [2, :]. The second row at any column?

Collectives™ on Stack Overflow

Extracting specific columns in numpy array by condition

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