numpy.unique


此函数返回输入数组中的唯一元素数组。该函数可以返回唯一值数组的数组和关联索引数组。索引的性质取决于函数调用中返回参数的类型。

numpy.unique(arr, return_index, return_inverse, return_counts)

这里

序号 参数和描述
1 arr 输入数组。如果不是一维阵列,将会变平
2 return_index 如果为True,则返回输入数组中元素的索引
3 return_inverse 如果为True,则返回唯一数组的索引,该索引可用于重建输入数组
4 return_counts 如果为True,则返回唯一数组中的元素出现在原始数组中的次数

import numpy as np
a = np.array([5,2,6,2,7,5,6,8,2,9])

print 'First array:'
print a
print '\n'  

print 'Unique values of first array:'
u = np.unique(a)
print u
print '\n'  

print 'Unique array and Indices array:'
u,indices = np.unique(a, return_index = True)
print indices
print '\n'  

print 'We can see each number corresponds to index in original array:'
print a
print '\n'  

print 'Indices of unique array:'
u,indices = np.unique(a,return_inverse = True)
print u
print '\n'

print 'Indices are:'
print indices
print '\n'  

print 'Reconstruct the original array using indices:'
print u[indices]
print '\n'  

print 'Return the count of repetitions of unique elements:'
u,indices = np.unique(a,return_counts = True)
print u
print indices

其输出如下

First array:
[5 2 6 2 7 5 6 8 2 9]

Unique values of first array:
[2 5 6 7 8 9]

Unique array and Indices array:
[1 0 2 4 7 9]

We can see each number corresponds to index in original array:
[5 2 6 2 7 5 6 8 2 9]

Indices of unique array:
[2 5 6 7 8 9]

Indices are:
[1 0 2 0 3 1 2 4 0 5]

Reconstruct the original array using indices:
[5 2 6 2 7 5 6 8 2 9]

Return the count of repetitions of unique elements:
[2 5 6 7 8 9]
 [3 2 2 1 1 1]