熊猫数据框，获取每行的最近日期

Question

I'm trying to go through each row (iterrow?) and find the most recent date (sort function?) and put it in column 'G'

I'm having trouble combining the iterate function and the sorting function.

    A       B           C           D           E           F           G
0   1       20171018    20171019    20171001    20171002    id_123      
1   2       NaN         20171005    20171006    20171003    id_234      
2   3       NaN         NaN         20171019    20171020    id_345      
3   4       NaN         NaN         NaN         20171021    id_456      

Desired Output

    A       B           C           D           E           F           G
0   1       20171018    20171019    20171001    20171002    id_123      20171019
1   2       NaN         20171005    20171006    20171003    id_234      20171006
2   3       NaN         NaN         20171019    20171020    id_345      20171020
3   4       NaN         NaN         NaN         20171021    id_456      20171021

Here is the code to generate the dataframe

data2 = {'A': [1, 2, 3, 4], 
        'B': ['20171018', '', '', ''], 
        'C': ['20171019', '20171005', '', ''],
        'D': ['20171001', '20171006', '20171019', ''],
        'E': ['20171002', '20171003', '20171020', '20171021'],
        'F': ['id_123','id_234','id_345','id_456'],
        'G': ['','','','']
        }
df3 = pd.DataFrame(data2)

edit: I have already converted the date columns using datetime

Answer 1

You can use the .max() method on the dataframe to get the most recent date. You will need to pass the parameter axis=1 to have it calculate the max along each row.

import pandas as pd

data = {'A': [1, 2, 3, 4],
        'B': ['20171018', '', '', ''],
        'C': ['20171019', '20171005', '', ''],
        'D': ['20171001', '20171006', '20171019', ''],
        'E': ['20171002', '20171003', '20171020', '20171021'],
        'F': ['id_123','id_234','id_345','id_456']
        }
df = pd.DataFrame(data)

# convert to datetimes
for c in 'BCDE':
    df[c] = pd.to_datetime(df[c])

# create a new column
df['G'] = df[['B','C','D','E']].max(axis=1)
print(df)

   A          B          C          D          E       F          G
0  1 2017-10-18 2017-10-19 2017-10-01 2017-10-02  id_123 2017-10-19
1  2        NaT 2017-10-05 2017-10-06 2017-10-03  id_234 2017-10-06
2  3        NaT        NaT 2017-10-19 2017-10-20  id_345 2017-10-20
3  4        NaT        NaT        NaT 2017-10-21  id_456 2017-10-21