根据 B 列的最大值替换 A 列的值

Question

Input:

ID  number  account date
1   00002   GA1     1/1/2021
1   00004   GA1     1/3/2021
2   00001   GA1     1/4/2021
3   00012   GA2     1/3/2021
4   00010   GA2     1/2/2021

Output:

ID  number  account date
1   00002   GA1     1/1/2021
1   00002   GA1     1/1/2021
1   00002   GA1     1/1/2021
4   00010   GA2     1/2/2021
4   00010   GA2     1/2/2021

Rules:

1. For the same account, replace all dates with the earliest date of that account
2. For the same account, replace all ID and number with the ID and number where the earliest date appears

I have done the first part of the query -

df.groupby(['account'])['date'].transform('min')

However, I am having trouble transform the ID and number columns based on the date column. I'm looking for an efficient way to do this.

Answer 1

Sorting by account and date then taking the first row of each group would give what you require.

The next step is to join it back to the original DataFrame

In [18]: df[['account']].join(df.sort_values(['account', 'date']).groupby('account').first(), on='account')
Out[18]:
  account  ID  number      date
0     GA1   1       2  1/1/2021
1     GA1   1       2  1/1/2021
2     GA1   1       2  1/1/2021
3     GA2   4      10  1/2/2021
4     GA2   4      10  1/2/2021