How to delete multiple rows of NumPy array?

Question

How can I delete multiple rows of NumPy array? For example, I want to delete the first five rows of x. I'm trying the following code:

import numpy as np
x = np.random.rand(10, 5)
np.delete(x, (0:5), axis=0)

but it doesn't work:

np.delete(x, (0:5), axis=0)
               ^
SyntaxError: invalid syntax

Answer 1

There are several ways to delete rows from NumPy array.

The easiest one is to use basic indexing as with standard Python lists:

>>> import numpy as np
>>> x = np.arange(35).reshape(7, 5)
>>> x
array([[ 0,  1,  2,  3,  4],
       [ 5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14],
       [15, 16, 17, 18, 19],
       [20, 21, 22, 23, 24],
       [25, 26, 27, 28, 29],
       [30, 31, 32, 33, 34]])
>>> result = x[5:]
>>> result
array([[25, 26, 27, 28, 29],
       [30, 31, 32, 33, 34]])

You can select not only rows but columns as well:

>>> x[:2, 1:4]
array([[1, 2, 3],
       [6, 7, 8]])

---

Another way is to use "fancy indexing" (indexing arrays using arrays):

>>> x[[0, 2, 6]]
array([[ 0,  1,  2,  3,  4],
       [10, 11, 12, 13, 14],
       [30, 31, 32, 33, 34]])

You can achieve the same using np.take:

>>> np.take(x, [0, 2, 6], axis=0)
array([[ 0,  1,  2,  3,  4],
       [10, 11, 12, 13, 14],
       [30, 31, 32, 33, 34]])

---

Yet another option is to use np.delete as in the question. For selecting the rows/columns for deletion it can accept slice objects, int, or array of ints:

>>> np.delete(x, slice(0, 5), axis=0)
array([[25, 26, 27, 28, 29],
       [30, 31, 32, 33, 34]])
>>> np.delete(x, [0, 2, 3], axis=0)
array([[ 5,  6,  7,  8,  9],
       [20, 21, 22, 23, 24],
       [25, 26, 27, 28, 29],
       [30, 31, 32, 33, 34]])

But all this time that I've been using NumPy I never needed this np.delete, as in this case it's much more convenient to use boolean indexing.

As an example, if I would want to remove/select those rows that start with a value greater than 12, I would do:

>>> mask_array = x[:, 0] < 12  # comparing values of the first column
>>> mask_array
array([ True,  True,  True, False, False, False, False])
>>> x[mask_array]
array([[ 0,  1,  2,  3,  4],
       [ 5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14]])
>>> x[~mask_array]  # ~ is an element-wise inversion
array([[15, 16, 17, 18, 19],
       [20, 21, 22, 23, 24],
       [25, 26, 27, 28, 29],
       [30, 31, 32, 33, 34]])

---

For more information refer to the documentation on indexing: https://docs.scipy.org/doc/numpy/reference/arrays.indexing.html

Answer 2

If you want to delete selected rows you can write like

np.delete(x, (1,2,5), axis = 0)

This will delete 1,2 and 5 th line, and if you want to delete like (1:5) try this one

np.delete(x, np.s_[0:5], axis = 0)

by this you can delete 0 to 4 lines from your array.

np.s_[0:5] --->> slice(0, 5, None) both are same.