Whether to return the indexes that can be used to reconstruct our input array. Check the example below for clarification. By default, return_inverse=False.

4. return_countslink | boolean | optional

Whether to return the number of occurrences of each value. By default, return_counts=False.

5. axislink | int or None | optional

The axis along which to look for unique values.

Axis	Meaning
0	Find unique rows
1	Find unique columns
None	Find unique values

By default, axis=None.

Return value

A Numpy array that contains the unique values of the input value. You will also get additional arrays depending on whether you flag any of the return_ parameters as True.

Examples

Basic usage

1D case

To find all unique values in a 1D array:


        
        
            
                
                
                    a = np.array([4,5,6,5])
np.unique(a)
                
            
            array([4, 5, 6])

2D case

Consider the following 2D array:


        
        
            
                
                
                    a = np.array([[4,5],[6,5]])
a
                
            
            array([[4, 5],
       [6, 5]])

To find all unique values in 2D array:


        
        
            
                
                
                    np.unique(a)
                
            
            array([4, 5, 6])

Getting the indices of the unique values

To get the indices of the unique values:


        
        
            
                
                
                    a = np.array([4,5,6,5])
arr_unique_values, arr_index = np.unique(a, return_index=True)
print(arr_unique_values)
print(arr_index)
                
            
            [4 5 6]
[0 1 2]

Getting the inverse indexes

To get the inverse indexes:


        
        
            
                
                
                    a = np.array([4,5,6,5])
arr_unique_values, arr_inverse = np.unique(a, return_inverse=True)
print(arr_unique_values)
print(arr_inverse)
                
            
            [4 5 6]
[0 1 2 1]

Here, the inverse can be used to reconstruct the original values:


        
        
            
                
                
                    arr_unique_values[arr_inverse]
                
            
            array([4, 5, 6, 5])

Getting the counts

To get the counts, set return_counts=True:


        
        
            
                
                
                    a = np.array([4,5,6,5])
arr_unique_values, arr_counts = np.unique(a, return_counts=True)
print(arr_unique_values)
print(arr_counts)
                
            
            [4 5 6]
[1 2 1]

Here, we see that the value 5 occurs twice in the original array.

Finding unique rows

Consider the following 2D array:


        
        
            
                
                
                    a = np.array([[4,5],[6,5],[4,5]])
a
                
            
            array([[4, 5],
       [6, 5],
       [4, 5]])

Here, we see that rows at index 0 and index 2 are duplicates. To get all the unique rows:


        
        
            
                
                
                    np.unique(a, axis=0)
                
            
            array([[4, 5],
       [6, 5]])

Finding unique columns

Consider the following 2D array:


        
        
            
                
                
                    a = np.array([[4,5,4],[6,8,6]])
a
                
            
            array([[4, 5, 4],
       [6, 8, 6]])

Here, we see that columns at index 0 and index 2 are duplicate. To get all unique columns:


        
        
            
                
                
                    np.unique(a, axis=1)
                
            
            array([[4, 5],
       [6, 8]])

Published by Isshin Inada

Edited by 0 others

Did you find this page useful?

thumb_up

thumb_down

Comment

Citation

Ask a question or leave a feedback...

thumb_up

thumb_down

chat_bubble_outline

settings

Enjoy our search

Hit / to insta-search docs and recipes!