NumPy Data Types

NumPy supports a much greater variety of numerical types than Python does. It essentially corresponds to the C language's data types, with some types corresponding to Python's built-in types. The table below lists common basic NumPy types.

NumPy's numeric types are instances of dtype objects, which are uniquely identified by a character, including np.bool_, np.int32, np.float32, etc.

Data Type Object (dtype)

A data type object (an instance of numpy.dtype class) describes how the bytes in the fixed-size block of memory corresponding to an array item should be interpreted. It describes the following aspects of the data:

• Type of the data (integer, float, or Python object)
• Size of the data (e.g., how many bytes are used to store an integer)
• Byte order of the data (little-endian or big-endian)
• In the case of structured types, the names of the fields, the data type of each field, and part of the memory block each field takes
• If the data type is a sub-array, its shape and data type

Byte order is determined by prefixing the data type with '<' or '>'. '<' means little-endian (least significant byte stored first). '>' means big-endian (most significant byte stored first).

A dtype object is constructed using the following syntax:

numpy.dtype(object, align, copy)

• object - The object to convert to a data type object
• align - If true, pads the fields to make it similar to C structs
• copy - Creates a copy of the dtype object. If false, it's a reference to the built-in data type object

Example

Let's understand through examples.

Example 1

import numpy as np
# Using scalar type
dt = np.dtype(np.int32)
print(dt)

Output:

int32

Example 2

import numpy as np
# int8, int16, int32, int64 can be replaced by equivalent string 'i1', 'i2', 'i4', 'i8'
dt = np.dtype('i4')
print(dt)

Output:

int32

Example 3

import numpy as np
# Byte order mark
dt = np.dtype('<i4')
print(dt)

Output:

int32

The following example demonstrates the use of a structured data type, where the type fields and corresponding actual types are created.

Example 4

# First create a structured data type
import numpy as np
dt = np.dtype([('age',np.int8)]) 
print(dt)

Output:

[('age', 'i1')]

Example 5

# Apply the data type to an ndarray object
import numpy as np
dt = np.dtype([('age',np.int8)]) 
a = np.array([(10,),(20,),(30,)], dtype = dt) 
print(a)

Output:

[(10,) (20,) (30,)]

dt = np.dtype([('age', np.int8)])
a = np.array([(10,), (20,), (30,)], dtype=dt)
print(a)

Output:

[(10,) (20,) (30,)]

Example 6

# Field names can be used to access the actual age column
import numpy as np
dt = np.dtype([('age', np.int8)])
a = np.array([(10,), (20,), (30,)], dtype=dt)
print(a['age'])

Output:

[10 20 30]

The following example defines a structured data type called student, which includes a string field 'name', an integer field 'age', and a float field 'marks', and applies this dtype to an ndarray object.

Example 7

import numpy as np
student = np.dtype([('name', 'S20'), ('age', 'i1'), ('marks', 'f4')])
print(student)

Output:

[('name', 'S20'), ('age', 'i1'), ('marks', 'f4')]

Example 8

import numpy as np
student = np.dtype([('name', 'S20'), ('age', 'i1'), ('marks', 'f4')])
a = np.array([('abc', 21, 50), ('xyz', 18, 75)], dtype=student)
print(a)

Output:

[('abc', 21, 50.0), ('xyz', 18, 75.0)]

Each built-in type has a unique character code that defines it, as follows: