Change column value based on another column's first characters in pandas-CodePudding

Have two columns start_date and end_date (strings):

Some of the rows in end_date are formated like 9999-12-12 i want to replace 9999 for the YYYY in start_date

I tried extracting the first 4 characters (got a pd.Series) for the start_date column and using that to replace but i cannot find a way

How it is

start_date	end_date
2020-12-25	2020-12-28
2021-02-02	9999-02-09
2019-02-13	9999-02-15

How it should be

start_date	end_date
2020-12-25	2020-12-28
2021-02-02	2021-02-09
2019-02-13	2019-02-15

CodePudding user response：

Assuming your data frame is called df:

df['end_date'] = df.apply(lambda row: row['end_date'].replace("9999", row['start_date'][0:4]), axis=1)

CodePudding user response：

Or use np.where:

df['end_date'] = np.where(df.end_date.str[:4] == '9999', df.start_date.str[:4]   df.end_date.str[4:], df.end_date)

df
   start_date    end_date
0  2020-12-25  2020-12-28
1  2021-02-02  2021-02-09
2  2019-02-13  2019-02-15