Test whether similar data types of different sizes are not subdtypes of each other in Python

The numpy.issubdtype() method in Python NumPy tests whether one data type is a subtype of another. When checking similar data types of different sizes (like float32 vs float64), they are not considered subtypes of each other despite being related.

Syntax

numpy.issubdtype(arg1, arg2)

Parameters

arg1, arg2: Data types or objects coercible to data types to compare for subtype relationship.

Import Required Library

First, import the NumPy library ?

import numpy as np

Testing Float Data Types

Check whether different float sizes are subtypes of each other ?

import numpy as np

print("Checking float32 vs float64:")
print("float32 subtype of float64:", np.issubdtype(np.float32, np.float64))
print("float64 subtype of float32:", np.issubdtype(np.float64, np.float32))

Checking float32 vs float64:
float32 subtype of float64: False
float64 subtype of float32: False

Testing Integer Data Types

Check whether different integer sizes are subtypes of each other ?

import numpy as np

print("Checking integer data types:")
print("int16 subtype of int32:", np.issubdtype(np.int16, np.int32))
print("int32 subtype of int16:", np.issubdtype(np.int32, np.int16))
print("int64 subtype of int32:", np.issubdtype(np.int64, np.int32))
print("int32 subtype of int64:", np.issubdtype(np.int32, np.int64))

Checking integer data types:
int16 subtype of int32: False
int32 subtype of int16: False
int64 subtype of int32: False
int32 subtype of int64: False

Complete Example

import numpy as np

print("Testing NumPy data type relationships\n")

# Float data types
print("Float data type comparisons:")
print("float32 vs float64:", np.issubdtype(np.float32, np.float64))
print("float64 vs float32:", np.issubdtype(np.float64, np.float32))

print("\nInteger data type comparisons:")
print("int16 vs int32:", np.issubdtype(np.int16, np.int32))
print("int32 vs int16:", np.issubdtype(np.int32, np.int16))
print("int64 vs int32:", np.issubdtype(np.int64, np.int32))
print("int32 vs int64:", np.issubdtype(np.int32, np.int64))

# Show actual parent types
print("\nActual parent types:")
print("float32 parent:", np.issubdtype(np.float32, np.floating))
print("int32 parent:", np.issubdtype(np.int32, np.integer))

Testing NumPy data type relationships

Float data type comparisons:
float32 vs float64: False
float64 vs float32: False

Integer data type comparisons:
int16 vs int32: False
int32 vs int16: False
int64 vs int32: False
int32 vs int64: False

Actual parent types:
float32 parent: True
int32 parent: True

Key Points

? Similar data types of different sizes (e.g., int16 and int32) are not subtypes of each other

? Each specific size represents a distinct data type in NumPy's type hierarchy

? All integer types are subtypes of np.integer, and all float types are subtypes of np.floating

Conclusion

NumPy treats data types of different sizes as separate, non-hierarchical types. Use issubdtype() with generic parent types like np.integer or np.floating to check broader type categories.