Convert 2D float array to 2D int array in NumPy

Converting a 2D float array to a 2D integer array in NumPy is a straightforward process using the astype() method. This conversion can be useful in various data analysis and scientific computing tasks where integer data types are required or where memory efficiency is essential. In this article, we are going to have a look at 'How to convert a 2D float array to a 2D int array in NumPy?'

What are 2D arrays in Python?

2D arrays mean '2-Dimensional arrays'. These are the arrays that contain rows and columns, thus making them 2-dimensional.

Converting 2D float array to 2D int array in NumPy

Below are the topics that we will cover in this article:

• Using astype() method
• Setting the dtype of the array to np.int32
• Using ndarray.round() method.
• Using np.floor(ndarray) function.
• Using np.ceil(ndarray) function.
• Using np.trunc(ndarray) function.
• Type-casting the array directly to the np.int32 data-type.

Using astype() method

To convert 2D float array to int array, we can make use of the astype() method.

Syntax: integer_array = integer_array.astype(np.data_type)

Example: In the program, we have used astype() to convert the float array to int array. Initially, the data type of the array was 'float'. In the astype() method, we gave the argument as np.int32 (integer data type) which changed the data type of the array.

import numpy as np

#  creating 2D float array. 
float_array = np.array([[1., 2., 3.], [4., 5., 6.]])
print('Before converting: ', float_array, '\n Data type of array: ', float_array.dtype)

# using astype() method to convert the data type of float_array from float to int
float_array = float_array.astype(np.int32)
print('After Converting: ', float_array, '\n Data type of array: ', float_array.dtype)

Output

Before converting:  [[1. 2. 3.]
 [4. 5. 6.]] 
 Data type of array:  float64
After Converting:  [[1 2 3]
 [4 5 6]] 
 Data type of array:  int32

Setting the dtype of the array to np.int32

When we initialize an array with float elements, its dtype (data-type) is np.float64. To convert the data-type of the array to int, we can change the dtype to np.int32.

import numpy as np

#  creating 2D float array. 
float_array = np.array([[1., 2., 3.], [4., 5., 6.]])
print('Before converting: ', float_array, '\n Data type of array: ', float_array.dtype)

# setting dtype to np.int32 to convert the data type of float_array from float to int
float_array = np.array(float_array, dtype=np.int32)
print('After Converting: ', float_array, '\n Data type of array: ', float_array.dtype)

Output:

Before converting:  [[1. 2. 3.]
 [4. 5. 6.]] 
 Data type of array:  float64
After Converting:  [[1 2 3]
 [4 5 6]] 
 Data type of array:  int32

In the above example, first, we created a 2D array with float elements. Then, we stored another array in the same variable (float_array) with the same elements by passing the float_array as argument in the np.array(). Finally, we changed the dtype of the array to np.int32. Thus, the 2D float array converted to 2D int array.

Using ndarray.round() method.

To convert the 2D float array to a 2D int array, we can first round the floating-point numbers to the nearest integer. This gives us a 2D float array as well but with rounded values. Then, we can finally convert it to the integer data-type.

# importing numpy library
import numpy as np

# creating a 2D float array
arr_2dim = np.array([1.35, 2.75, 3.50, 4.0, 5.9, 6.85]).reshape(2, 3)
print('Before converting:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

print()

# Rounding the array
arr_2dim = arr_2dim.round()
print('After rounding:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

print()

# After rounding to the nearest integer, the data-type of the array is still float.
# Therefore, to convert it, we can type-cast.
arr_2dim = np.int32(arr_2dim)
print('After converting:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

Output:

Before converting:
 [[1.35 2.75 3.5 ]
 [4.   5.9  6.85]] 
Data-type of the array:  float64
After rounding:
 [[1. 3. 4.]
 [4. 6. 7.]] 
Data-type of the array:  float64
After converting:
 [[1 3 4]
 [4 6 7]] 
Data-type of the array:  int32

In the above program, we first created a 2D float array. Then, we rounded it. Inside the round function, the default value of decimals to be rounded is 'zero'. That is why, it rounded in the manner that there are no numbers but zero after the decimal point. But still, even after the rounding, the data-type of the array is float. Thus, we type-casted the array into integer data-type.

In the place of round function, we can also use 'np.round_(ndarray)' function, which is nothing but the alias of the round function. It provides the exact same functionality as the round function.

Using np.floor(ndarray) function.

We can use np.floor(ndarray) to floor the elements in the array and then we can later type-cast it to the integer data-type.

np.floor(ndarray) : This function floors the elements in the array, i.e., converts the numbers to the nearest integer to them, but lesser than them.

# importing numpy library
import numpy as np

# creating a 2D float array
arr_2dim = np.array([1.35, 2.75, 3.50, 4.0, 5.9, 6.85]).reshape(2, 3)
print('Before converting:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

print()

# using the floor function on the array
arr_2dim = np.floor(arr_2dim)
print('After using the floor function:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

print()

# After flooring to the nearest, smaller integer, the data-type of the array is still float.
# Therefore, to convert it, we can type-cast.
arr_2dim = np.int32(arr_2dim)
print('After converting:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

Output:

Before converting:
 [[1.35 2.75 3.5 ]
 [4.   5.9  6.85]] 
Data-type of the array:  float64
After using the floor function:
 [[1. 2. 3.]
 [4. 5. 6.]] 
Data-type of the array:  float64
After converting:
 [[1 2 3]
 [4 5 6]] 
Data-type of the array:  int32

Using np.ceil(ndarray) function.

np.ceil(ndarray) : This function ceils the elements in the array, i.e., convert the numbers to the nearest integer to them, but greater than them.

We can, first, ceil the elements in the array and then later convert it to the integer data-type.

# importing numpy library
import numpy as np

# creating a 2D float array
arr_2dim = np.array([1.35, 2.75, 3.50, 4.0, 5.9, 6.85]).reshape(2, 3)
print('Before converting:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

print()

# using the ceil function on the array
arr_2dim = np.ceil(arr_2dim)
print('After using the ceil function:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

print()

# After ceiling to the nearest, greater integer, the data-type of the array is still float.
# Therefore, to convert it, we can type-cast.
arr_2dim = np.int32(arr_2dim)
print('After converting:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

Output:

Before converting:
 [[1.35 2.75 3.5 ]
 [4.   5.9  6.85]] 
Data-type of the array:  float64
After using the ceil function:
 [[2. 3. 4.]
 [4. 6. 7.]] 
Data-type of the array:  float64
After converting:
 [[2 3 4]
 [4 6 7]] 
Data-type of the array:  int32

Using the np.trunc(ndarray) function.

np.trunc(ndarray) : This function truncates the elements in the array, i.e., converts them to the integer that is closer to the zero than the element itself is. In other words, just the the numbers after the decimal point are removed.

# importing numpy library
import numpy as np

# creating a 2D float array
arr_2dim = np.array([1.35, 2.75, 3.50, 4.0, 5.9, 6.85]).reshape(2, 3)
print('Before converting:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

print()

# using the truncate method on the array
arr_2dim = np.trunc(arr_2dim)
print('After using the trunc function:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

print()

# After truncating to the nearest integer to the zero than the element is, the data-type of the array is still float.
# Therefore, to convert it, we can type-cast.
arr_2dim = np.int32(arr_2dim)
print('After converting:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

Output:

Before converting:
 [[1.35 2.75 3.5 ]
 [4.   5.9  6.85]] 
Data-type of the array:  float64
After using the trunc function:
 [[1. 2. 3.]
 [4. 5. 6.]] 
Data-type of the array:  float64
After converting:
 [[1 2 3]
 [4 5 6]] 
Data-type of the array:  int32

In the place of np.trunc() function, we can also use np.fix().

In all the above examples, we have used np.int32() to convert the data-type. We could also use astype() method for the same.

Type-casting the array directly to the np.int32 data-type.

We can also convert the elements of an array to the integer data-type by type-casting as given below:

ndarray = np.int32(ndarray)

# importing numpy library
import numpy as np

# creating a 2D float array
arr_2dim = np.array([1.35, 2.75, 3.50, 4.0, 5.9, 6.85]).reshape(2, 3)
print('Before converting:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

print()

# After type-casting.
arr_2dim = np.int32(arr_2dim)
print('After converting:\n', arr_2dim, '\nData-type of the array: ', arr_2dim.dtype)

Output:

Before converting:
 [[1.35 2.75 3.5 ]
 [4.   5.9  6.85]] 
Data-type of the array:  float64
After converting:
 [[1 2 3]
 [4 5 6]] 
Data-type of the array:  int32

We can convert the 2D int array to 2D float array in the same manner.