NumPy: numpy.unique() function

numpy.unique() function

The numpy.unique() function is used to find the unique elements of an array.Returns the sorted unique elements of an array. There are three optional outputs in addition to the unique elements:

• the indices of the input array that give the unique values
• the indices of the unique array that reconstruct the input array
• the number of times each unique value comes up in the input array

This function is useful when we need to identify the unique values of an array and use them for further processing, such as counting occurrences or filtering out duplicates. It is commonly used in data analysis and machine learning applications, as well as in general programming tasks.

Syntax:

numpy.unique(filt, trim='fb')

Parameters:

Return value:

unique : ndarray - The sorted unique values.
unique_indices : ndarray, optional - The indices of the first occurrences of the unique values in the original array. Only provided if return_index is True.
unique_inverse : ndarray, optional - The indices to reconstruct the original array from the unique array. Only provided if return_inverse is True.
unique_counts : ndarray, optional - The number of times each of the unique values comes up in the original array. Only provided if return_counts is True.

Example: Finding unique elements in a numpy array

>>> import numpy as np
>>> np.unique([0,1,2,0,2,3,4,3,0,4])
array([0, 1, 2, 3, 4])

In the above code, the input array is [0,1,2,0,2,3,4,3,0,4]. The numpy.unique() function returns a sorted array of unique elements present in the input array, in ascending order.

Pictorial Presentation:

Example: Finding unique elements in a 2D array using numpy.unique()

>>> import numpy as np
>>> x = np.array([[1, 1], [2,3], [3,4]])
>>> np.unique(x)
array([1, 2, 3, 4])

In the above code we first create a 2D array 'x' using the np.array() method with the values [[1, 1], [2, 3], [3, 4]]. Then, the np.unique() function is used to find the unique elements in the array 'x'.
Since 'x' is a 2D array, np.unique() treats it as a flattened array and returns only the unique values. Therefore, the output is an array of unique elements [1, 2, 3, 4].

Pictorial Presentation:

Example: Finding unique elements and their indices in a numpy array

>>> import numpy as np
>>> x = np.array(['o', 'p', 'y', 't', 'h', 'o', 'p'])
>>> u, indices = np.unique(x, return_index=True)
>>> u
array(['h', 'o', 'p', 't', 'y'], 
      dtype='<U1')
>>> indices
array([4, 0, 1, 3, 2])
>>> x[indices]
array(['h', 'o', 'p', 't', 'y'], 
      dtype='<U1')

The above code defines a numpy array x consisting of the string values ['o', 'p', 'y', 't', 'h', 'o', 'p']. The np.unique() function is called on the array x with the parameter return_index=True to get unique elements and their indices. The returned values are assigned to u and indices respectively. The u array consists of the unique elements of the x array in sorted order, i.e. ['h', 'o', 'p', 't', 'y']. The indices array consists of the indices of the first occurrences of these unique elements in the x array, i.e. [4, 0, 1, 3, 2]. Finally, the code uses the indices array to retrieve the original elements in the same order as in the u array, i.e. ['h', 'o', 'p', 't', 'y'], by accessing the corresponding elements in the x array.

Example: Finding unique values and their indices using numpy.unique() function with return_inverse parameter

>>> import numpy as np
>>> x = np.array([0, 1, 2, 5, 2, 6, 5, 2, 3, 1])
>>> u, indices = np.unique(x, return_inverse=True)
>>> u
array([0, 1, 2, 3, 5, 6])

In the above code the np.unique() function takes an array as input and returns an array of unique elements in that array. In this case, the function takes the x array as input and returns an array u that contains the unique elements in x.
The second parameter return_inverse is set to True. This will also return an array of indices that can be used to reconstruct the original array from the unique values. The indices array indices is returned.

Pictorial Presentation:

Example: Mapping an array using numpy.unique()

>>> import numpy as np
>>> x = np.array([0, 1, 2, 5, 2, 6, 5, 2, 3, 1])
>>> indices
array([0, 1, 2, 4, 2, 5, 4, 2, 3, 1])
>>> u[indices]
array([0, 1, 2, 5, 2, 6, 5, 2, 3, 1])

In the above code:

• In the first line, we create an array 'x' with 10 elements. Then, we use the numpy.unique() function with the argument 'return_inverse=True'. This returns two outputs - the unique values in the array 'u' and an array 'indices' which contains the indices of the unique values corresponding to each element in the input array.
• In the second line, we print the unique values obtained from the first output.
• In the third line, we print the 'indices' array, which contains the indices of the unique values corresponding to each element in the input array.
• In the fourth line, we use the 'indices' array to map each element in the input array 'x' to its corresponding unique value from the 'u' array. The resulting mapped array is printed in the output.

Example: Finding unique values in an array using numpy.unique() with return_inverse=True

>>> import numpy as np
>>> x = np.array([1,2,5,3,4,2,3,2,5,4])
>>> u, indices = np.unique(x, return_inverse=True)
>>> u
array([1, 2, 3, 4, 5])

The above code uses the NumPy function np.unique() to find the unique values in a NumPy array 'x'. The function is called with the optional argument 'return_inverse=True', which also returns an array of indices of the unique values in the original array.
The unique values are returned as an array 'u', which is sorted in ascending order. In this case, the unique values in 'x' are [1, 2, 3, 4, 5]. The array 'indices' is also returned, which represents the original array 'x' in terms of the unique values in 'u'.

Pictorial Presentation:

Example: Mapping original values using unique and inverse index in numpy

>>> import numpy as np
>>> x = np.array([1,2,5,3,4,2,3,2,5,4])
>>> u, indices = np.unique(x, return_inverse=True)
>>> indices
array([0, 1, 4, 2, 3, 1, 2, 1, 4, 3])
>>> u[indices]
array([1, 2, 5, 3, 4, 2, 3, 2, 5, 4])

In the above code, first a NumPy array x is created with some repeated values. Then, the np.unique function is used to find the unique elements in the array x and the indices that reconstruct the original array from unique values. The returned unique values are stored in the u variable and the indices that reconstruct the original array are stored in the indices variable.
The indices variable is an array of indices, which are the positions of each element in the u array that corresponds to the corresponding element in the original x array. The u[indices] expression is then used to map the original x values back from the unique values and indices.
The resulting array is the same as the original array x, but with each element mapped to the corresponding unique value.

Python - NumPy Code Editor:

Previous: trim_zeros()
Next: Rearrangeing elements flip()