Question

what is the simplest way to get all rows where a complex condition holds for an ndarray that represents a 2d matrix? e.g. get all rows where all the values are above 5 or all the values are below 5?

thanks.

Answer 1

You probably know that boolean arrays can be used for indexing, e.g.:

import numpy as np
x = np.arange(10)
x2 = x[x<5]

For a boolean array, np.all lets you apply it across a given axis:

y = np.arange(12).reshape(3,4)

b = y < 6
b1 = np.all(b, axis=0)
b2 = np.all(b, axis=1)

y1 = y[b1]
y2 = y[b2]

It only returns the arrays which meet the criteria, so the shape is not preserved. (If you do need to preserve the shape, then take a look at masked arrays.)