Python 列表到 np.array 的计数

Question

Suppose we have a vector size N=1000 and let's say we get the list [1,1,2,2,2,100]

I'd like to generate an np.array (or pd.Series) of size 1000 where v[n] is the number of times n appears in the list. In our example, v[1] = 2, v[2] = 3, v[100] = 1, v=[42] = 0

How can I do that with numpy/pandas elegantly?

Answer 1

If you have a list mylist , you can get an array of counts mycount :

N = 1000
x = np.array(mylist)
mycount = np.bincount(x, minlength=N)

This sorts each element in the array into bins based on its value and quantity. You can find more information on bincount on this doc page .

Answer 2

Python has a native method for counting occurrences called Counter which can be used without invoking numpy or pandas if desired

from collections import Counter
a = [1,1,2,2,2,100]
cnts = Counter(a)
print(cnts)
# Counter({2: 3, 1: 2, 100: 1})

You can convert this to a list with a list comprehension:

N = 100
cnts_list = [cnts.get(i, 0) for i in range(N+1)]

Answer 3

Use Series.value_counts with Series.reindex for add non exist values:

a = [1,1,2,2,2,100]

N = 100
a = pd.Series(a).value_counts().reindex(range(N+1), fill_value=0)
print (a)
0      0
1      2
2      3
3      0
4      0
      ..
96     0
97     0
98     0
99     0
100    1
Length: 101, dtype: int64

Answer 4

You can use np.unique as well.

N = 1000
result = np.zeros(N)
idx, val = np.unique([1,1,2,2,2,100], return_counts=True)
result[idx] = val
print(result[:5])                                                                                                                                                                                                                                                           
>>>[0. 2. 3. 0. 0.]

more information: https://numpy.org/doc/stable/reference/generated/numpy.unique.html

Answer 5

You can use Series and Group by

In[1]:

import pandas as pd
my_list = [1,1,1,2,2,2,2,3,4,8,1000,8,8,5,5,6]

my_Serie = pd.Series(my_list)
v = my_Serie.groupby(my_list).count().to_dict()
print(v)

{1: 3, 2: 4, 3: 1, 4: 1, 5: 2, 6: 1, 8: 3, 1000: 1}