Arrays used as indices must be of integer (or boolean) type

# Arrays used as indices must be of integer (or boolean) type

The NumPy "IndexError: arrays used as indices must be of integer (or boolean) type" occurs when you try to use an array containing non-integer or boolean values when indexing another array.

To solve the error, use the astype() method to convert the type of the array elements to integers before using it to index another array.

Here is an example of how the error occurs.

main.py

Copied!
import numpy as np

indices_array = np.array([
    [0, 1.1, 2],
    [1.3, 3.2, 2]
])

array = np.array([
    [1, 2, 3, 4],
    [5, 6, 7, 8],
    [9, 10, 11, 12],
    [13, 14, 15, 16]
])


indices_column = indices_array[:, 0]
print(indices_column)  # 👉️ [0.  1.3]

# ⛔️ IndexError: arrays used as indices must be of integer (or boolean) type
print(array[indices_column])

index error arrays used as indices must be of integer or boolean type

The indices_array variable stores an array that we will use the index the other array.

However, notice that some of the values in the indices array are of type float.

main.py

Copied!
import numpy as np

indices_array = np.array([
    [0, 1.1, 2],
    [1.3, 3.2, 2]
])

# [[0.  1.1 2. ]
#  [1.3 3.2 2. ]]
print(indices_array)

The code for this article is available on GitHub

You can use the dtype attribute to check the type of the values in the NumPy array.

main.py

Copied!
import numpy as np

indices_array = np.array([
    [0, 1.1, 2],
    [1.3, 3.2, 2]
])

print(indices_array.dtype) # 👉️ float64

Notice that the type of the values in the array is float64.

The values in the array have to either be integers or booleans for us to be able to use the array to index another array.

# Using the `astype()` method to solve the error

One way to solve the error is to use the astype() method to convert the values in the array to integers.

main.py

Copied!
import numpy as np

indices_array = np.array([
    [0, 1.1, 2],
    [1.3, 3.2, 2]
])


array = np.array([
    [1, 2, 3, 4],
    [5, 6, 7, 8],
    [9, 10, 11, 12],
    [13, 14, 15, 16]
])

indices_column = indices_array[:, 0].astype(int)
print(indices_column)  # 👉️ [0 1]

print(indices_column.dtype)  # 👉️ int64

# [[1 2 3 4]
#  [5 6 7 8]]
print(array[indices_column])

using astype method to solve the error

The code for this article is available on GitHub

We used array slicing to get the first column of the indices array and called the ndarray.astype() method on it.

The method copies the array and casts it to the specified type.

You can use the dtype attribute to verify that the indices array contains integer values.

main.py

Copied!
indices_column = indices_array[:, 0].astype(int)
print(indices_column)  # 👉️ [0 1]

print(indices_column.dtype)  # 👉️ int64

The values in the array have to either be of type int or bool.

If you convert the values in the indices array to booleans, 0 of any numeric type gets converted to False.

main.py

Copied!
import numpy as np

indices_array = np.array([
    [0, 1.1, 2],
    [1.3, 3.2, 2],
    [2.3, 3.2, 2],
    [3.3, 3.2, 2],
])


array = np.array([
    [1, 2, 3, 4],
    [5, 6, 7, 8],
    [9, 10, 11, 12],
    [13, 14, 15, 16]
])

indices_column = indices_array[:, 0].astype(bool)
print(indices_column)  # 👉️ [False  True  True  True]

print(indices_column.dtype)  # 👉️ bool

# [[ 5  6  7  8]
#  [ 9 10 11 12]
#  [13 14 15 16]]
print(array[indices_column])

using astype boolean

The code for this article is available on GitHub

When you pass the bool type to the astype method, only 0 (zero) values of any numeric type get converted to False, assuming you have a numeric array.

If your array contains values of other types, then all falsy values will get converted to False.

The falsy values in Python are:

constants defined to be falsy: None and False.
0 (zero) of any numeric type
empty sequences and collections: "" (empty string), () (empty tuple), [] (empty list), {} (empty dictionary), set() (empty set), range(0) (empty range).

All other values are truthy.

# Setting the `dtype` of the indices array to `int` upon initialization

You can also set the dtype of the indices array to int upon initialization to solve the error.

main.py

Copied!
import numpy as np

indices_array = np.array([
    [0, 1.1, 2],
    [1.3, 3.2, 2],
], dtype=int)


array = np.array([
    [1, 2, 3, 4],
    [5, 6, 7, 8],
    [9, 10, 11, 12],
    [13, 14, 15, 16]
])

indices_column = indices_array[:, 0]
print(indices_column)  # 👉️ [0 1]

print(indices_column.dtype)  # 👉️ int64

# [[1 2 3 4]
#  [5 6 7 8]]
print(array[indices_column])

set dtype to int upon initialization

The code for this article is available on GitHub

Notice that we set the dtype argument to int when creating the indices array.

The numpy.array() method takes an optional dtype argument.

The argument is used to specify the desired data type for the array.

If you don't explicitly supply the dtype argument, NumPy tries to use a default dtype that can represent the values in the array.

# Additional Resources

You can learn more about the related topics by checking out the following tutorials:

I wrote a book in which I share everything I know about how to become a better, more efficient programmer.

You can use the search field on my Home Page to filter through all of my articles.

Arrays used as indices must be of integer (or boolean) type