Return True if cast between data types can occur according to the casting rule in Python

The numpy.can_cast() method returns True if a cast between data types can occur according to NumPy's casting rules. It takes two parameters: the source data type and the target data type to cast to.

Syntax

numpy.can_cast(from_dtype, to_dtype, casting='safe')

Parameters

from_dtype: The data type or array to cast from
to_dtype: The data type to cast to
casting: Controls what kind of data casting may occur ('no', 'equiv', 'safe', 'same_kind', 'unsafe')

Basic Usage

Let's check if various data type conversions are possible ?

import numpy as np

print("Checking with can_cast() method in NumPy\n")
print("int32 to int64:", np.can_cast(np.int32, np.int64))
print("float64 to complex:", np.can_cast(np.float64, complex))
print("complex to float:", np.can_cast(complex, float))
print("int64 to float64:", np.can_cast('i8', 'f8'))
print("int64 to float32:", np.can_cast('i8', 'f4'))
print("int32 to string:", np.can_cast('i4', 'S4'))

Checking with can_cast() method in NumPy

int32 to int64: True
float64 to complex: True
complex to float: False
int64 to float64: True
int64 to float32: False
int32 to string: False

Understanding the Results

The method follows NumPy's casting hierarchy:

• int32 to int64: Safe upcast (smaller to larger integer)
• float64 to complex: Safe conversion (real numbers can become complex)
• complex to float: Unsafe (would lose imaginary part)
• int64 to float32: Unsafe (potential precision loss)

Using Different Casting Rules

import numpy as np

# Safe casting (default)
print("Safe casting:")
print("int64 to float32 (safe):", np.can_cast('i8', 'f4', casting='safe'))

# Same kind casting
print("\nSame kind casting:")
print("int64 to float32 (same_kind):", np.can_cast('i8', 'f4', casting='same_kind'))

# Unsafe casting
print("\nUnsafe casting:")
print("int64 to float32 (unsafe):", np.can_cast('i8', 'f4', casting='unsafe'))

Safe casting:
int64 to float32 (safe): False

Same kind casting:
int64 to float32 (same_kind): True

Unsafe casting:
int64 to float32 (unsafe): True

Conclusion

The numpy.can_cast() method helps determine if data type conversions are safe according to NumPy's casting rules. Use it to prevent data loss when converting between different numeric types in your arrays.