Understanding NumPy’s unique Function
The numpy.unique function serves a straightforward yet indispensable purpose: it returns a sorted array of unique elements from an input array. This is incredibly useful for data cleaning, analysis, and preprocessing tasks where identifying distinct values is essential. Beyond simply returning the unique elements, unique also provides optional outputs that reveal the index of the first occurrence of each unique element in the input array and the counts of each unique element.
Basic Usage: Extracting Unique Elements
The most basic application of numpy.unique involves obtaining a sorted list of unique values. Consider the following example:
import numpy as np
arr = np.array([1, 2, 2, 3, 4, 4, 4, 5, 1])
unique_elements = np.unique(arr)
print(unique_elements) # Output: [1 2 3 4 5]This code snippet demonstrates how easily we can extract the unique elements from a NumPy array. The output is a sorted array containing only the distinct values from the input array, eliminating duplicates.
Advanced Usage: Indices and Counts
numpy.unique offers more than just unique elements; it also provides information about their indices and counts. The function can return three outputs:
unique_elements: The sorted array of unique values.indices: The indices of the first occurrence of each unique element in the input array.counts: The number of times each unique element appears in the input array.
Let’s see this in action:
import numpy as np
arr = np.array([1, 2, 2, 3, 4, 4, 4, 5, 1])
unique_elements, indices, counts = np.unique(arr, return_index=True, return_counts=True)
print("Unique elements:", unique_elements) # Output: [1 2 3 4 5]
print("Indices:", indices) # Output: [0 1 3 4 7]
print("Counts:", counts) # Output: [2 2 1 3 1]Here, return_index=True and return_counts=True enable the function to return the indices and counts, respectively. This allows for more detailed analysis of the unique values within the array.
Handling Multi-dimensional Arrays
numpy.unique seamlessly handles multi-dimensional arrays as well. However, by default, it flattens the array before finding unique elements. If you need to find unique rows or columns, you need to specify the axis parameter:
import numpy as np
arr_2d = np.array([[1, 2], [1, 2], [3, 4]])
unique_rows = np.unique(arr_2d, axis=0)
print(unique_rows) # Output: [[1 2] [3 4]]This example shows how to find unique rows in a 2D array. Specify axis=1 to find unique columns.
Working with Different Data Types
The unique function works effectively with various data types, including strings and other custom data types supported by NumPy.
import numpy as np
arr_str = np.array(['apple', 'banana', 'apple', 'orange'])
unique_strings = np.unique(arr_str)
print(unique_strings) # Output: ['apple' 'banana' 'orange']This showcases the versatility of numpy.unique across different data types. Remember that for custom objects, appropriate comparison methods must be defined for accurate results.