NumPy ndarray.astype() Method

NumPy ndarray.astype()

The numpy.ndarray.astype() method is used to cast a NumPy array to a specified data type. It allows for explicit type conversion while providing control over memory order, casting rules, and whether to create a copy of the array.

Syntax

ndarray.astype(dtype, order='K', casting='unsafe', subok=True, copy=True)

Parameters

Parameter	Type	Description
dtype	data-type	The target data type to which the array elements should be converted.
order	{‘C’, ‘F’, ‘A’, ‘K’}, optional	Defines memory layout order: 'C' (row-major), 'F' (column-major), 'A' (match input), 'K' (match memory layout).
casting	{‘no’, ‘equiv’, ‘safe’, ‘same_kind’, ‘unsafe’}, optional	Defines casting rules, controlling whether conversions are allowed based on safety.
subok	bool, optional	If True, subclasses of ndarray are preserved; otherwise, a base ndarray is returned.
copy	bool, optional	If True, a copy of the array is always created; otherwise, it avoids copying when possible.

Return Value

Returns a new array with elements cast to the specified dtype. If copy=False and conversion is not required, the original array is returned.

---

Examples

1. Converting an Integer Array to Float

This example converts an integer array to a float array using astype().

import numpy as np

# Creating an integer array
arr = np.array([1, 2, 3, 4, 5])

# Converting to float type
float_arr = arr.astype(float)

# Printing original and converted arrays
print("Original array:", arr)
print("Converted array (float):", float_arr)

Output:

Original array: [1 2 3 4 5]
Converted array (float): [1. 2. 3. 4. 5.]

The integer values are successfully converted to floating-point numbers.

2. Converting a Float Array to Integer

This example demonstrates truncation when converting float values to integers.

import numpy as np

# Creating a float array
arr = np.array([1.9, 2.8, 3.7, 4.5])

# Converting to integer type
int_arr = arr.astype(int)

# Printing original and converted arrays
print("Original array:", arr)
print("Converted array (int):", int_arr)

Output:

Original array: [1.9 2.8 3.7 4.5]
Converted array (int): [1 2 3 4]

Note that floating-point values are truncated, not rounded, during conversion.

3. Converting an Array to Boolean

This example converts numerical values to boolean, where nonzero values become True and zero becomes False.

import numpy as np

# Creating an integer array
arr = np.array([0, 1, 2, 0, -3, 4])

# Converting to boolean type
bool_arr = arr.astype(bool)

# Printing original and converted arrays
print("Original array:", arr)
print("Converted array (boolean):", bool_arr)

Output:

Original array: [ 0  1  2  0 -3  4]
Converted array (boolean): [False  True  True False  True  True]

Zeros are converted to False, and all other values are converted to True.

4. Using casting Parameter in ndarray.astype()

The casting parameter controls the safety of type conversions.

import numpy as np

# Creating a float array
arr = np.array([1.2, 2.5, 3.8])

# Safe casting to integer (prevents unsafe conversions)
try:
    int_arr = arr.astype(int, casting='safe')
    print("Converted array:", int_arr)
except TypeError as e:
    print("Error:", e)

Output:

Converted array: [1 2 3]

The safe casting rule ensures that only safe conversions occur.