NumPy ndarray.getfield() Method

NumPy ndarray.getfield()

The numpy.ndarray.getfield() method extracts a specific field from a structured NumPy array. It allows retrieving values from an array that has a defined dtype with multiple fields.

Syntax

</>

Copy

ndarray.getfield(dtype, offset=0)

Parameters

Parameter	Type	Description
`dtype`	data-type	The data type of the field to be retrieved.
`offset`	int, optional	Byte offset from the start of the field, default is `0`.

Return Value

Returns a new array containing the extracted field values. The shape of the output remains the same as the original array, but only the selected field values are included.

Examples

1. Extracting a Specific Field from a Structured Array

In this example, we define a structured NumPy array with multiple fields and extract a specific field using getfield().

</>

Copy

import numpy as np

# Define a structured array with two fields: 'name' (string) and 'age' (integer)
dtype = np.dtype([('name', 'S10'), ('age', np.int32)])
data = np.array([('Alice', 25), ('Bob', 30), ('Charlie', 22)], dtype=dtype)

# Extract only the 'age' field using getfield
age_field = data.getfield(np.int32, offset=10)
print(age_field)

Output:

[25 30 22]

Here, getfield() retrieves only the integer field age, returning an array with the corresponding values.

2. Using `offset` to Extract Data from a Raw Byte Array

We can extract a field at a specific byte offset in a raw byte array using the offset parameter.

</>

Copy

import numpy as np

# Create a raw byte array with an arbitrary structure
raw_data = np.array([ (b'X', 100, 1.5),
                      (b'Y', 200, 2.5),
                      (b'Z', 150, 3.5)], 
                      dtype=[('char', 'S1'), ('int', np.int32), ('float', np.float64)])

# Extract the 'int' field with explicit offset from start (0 bytes)
int_field = raw_data.getfield(np.int32, offset=1)
print(int_field)

Output:

[100 200 150]

By specifying an offset, getfield() extracts the integer values starting from the correct memory location within the structured array.

3. Extracting Floating-Point Values from a Structured Array

Here, we extract a floating-point field from a structured array that contains multiple data types.

</>

Copy

import numpy as np

# Define a structured array with three fields: character, integer, and floating-point
dtype = np.dtype([('char', 'S1'), ('int', np.int32), ('float', np.float64)])
data = np.array([(b'A', 10, 3.14), 
                 (b'B', 20, 2.71), 
                 (b'C', 30, 1.61)], dtype=dtype)

# Extract the floating-point field
float_field = data.getfield(np.float64, offset=5)
print(float_field)

Output:

[3.14 2.71 1.61]

The getfield() method isolates the floating-point values from the structured array, producing an array of float values.

TutorialKart

NumPy ndarray.getfield() Method