使用pandas + python（有条件）在滚动窗口中计算不同的字符串

Question

I want to calculate the number of distinct port numbers that exist between the current row and the 5 previous rows (sliding window) and this when the same address appears. For instance,

If the input is (csv file):

ID      PORT     ADDRESS
1        21       ad3 
2        22       ad1  
3        23       ad2
4        25       ad2 
5        25       ad1
6        22       ad1 
7        22       ad1
8        21       ad4

The output should be:

ID      PORT     ADDRESS      COUNT_DISC_PORT
1        21       ad3        -
2        22       ad1        -
3        23       ad2        - 
4        25       ad2        - 
5        25       ad1        - 
6        22       ad1        2 
7        23       ad1        3
8        21       ad4        1 

I have read the documentation about the rolling function in pandas and I have tried combining group by and rolling with no success.

I am using Python 3.7 and the pandas package 0.22. Any feedback would be appreciated.

Answer 1

for index, row in df.iterrows(): small_df = df[index - 5:index] df['uniques'][index] = len(small_df.unique())

这是我的简要介绍。

Answer 2

Ok , seems like you data inout is mismatch with the df your show to us

df.groupby('ADDRESS').PORT.apply(lambda x : pd.Series(x).rolling(5,min_periods=1).apply(lambda y: len(set(y))))
Out[844]: 
0    1.0
1    1.0
2    1.0
3    2.0
4    2.0
5    2.0
Name: PORT, dtype: float64