I have an array in Python with 3 columns

You can use a simple list comprehension:

Get a list of all sub-lists with a number in the third index that is smaller than 5:

[row for row in arr if row[2] < 5]

Or if you want to get a list containing only the third number of each sub-list that is smaller than 5:

[row[2] for row in arr if row[2] < 5]

Use list comprehension to filter the third values less than 5.

import numpy as np

data = np.array([[ 44410,   2003,     18],
       [  5712,   1991,     30],
       [ 37123,   1990,     31],
       [     0,   2019,      2],
       [ 25757,   2006,     15],
       [ 10728,   2012,      9],
       [     0,   2019,      2],
       [ 77599,   2009,     12],
       [ 99197,   2010,     11]])

third_colum_filtered = [i[2] for i in data if i[2]<5]

print(third_colum_filtered)

Output:

[2, 2]

First you need to create a filter array:

>>> data[:,2] < 5
array([False, False, False,  True, False, False,  True, False, False], dtype=bool)>>> 

And then apply it to yours:

>>> filter_array = data[:,2] < 5
>>> data[filter_array]
array([[   0, 2019,    2],
       [   0, 2019,    2]])

So, short circuiting:

data[data[:,2] < 5]

Is this what you are after? If you want only the values then add ", 2" and that would convert the last column to a vector with the values:

>>> data[data[:,2] < 5, 2]
array([2, 2])