Group by unique Name and Status with the last Date

Question

I would like to analyze statistics per cars which were repairs and which are new. Data sample is:

Name   IsItNew    ControlDate
Car1    True      31/01/2018
Car2    True      28/02/2018
Car1    False     15/03/2018
Car2    True      16/04/2018
Car3    True      30/04/2018
Car2    False     25/05/2018
Car1    False     30/05/2018    

So, I should groupby by Name and if there is a False in IsItNew column I should set False and the first date, when False was happened.

I tried groupby with nunique():

df = df.groupby(['Name','IsItNew', 'ControlDate' ])['Name'].nunique()

But, it returns count of unique items in each group.

How can I receive only grouped unique items without any count?

Actual result is:

Name   IsItNew       ControlDate
Car1    True         31/01/2018     1
        False        15/03/2018     1
                     30/05/2018     1
Car2    True         28/02/2018     1
                     16/04/2018     1
        False        25/05/2018     1 
Car3    True         30/04/2018     1


Expected Result is:

Name   IsItNew     ControlDate
Car1    False      15/03/2018
Car2    False      25/05/2018
Car3    True       30/04/2018

I'd appreciate for any idea. Thanks)

Answer 1

One way to do it would be to GroupBy the Name, and aggregate on IsItNew with two functions. A custom one using any to check if there are any False values, and idxmin, to find the index of the first False, which you can later on use to index the dataframe on ControlDate:

df_ = df.groupby('Name').agg({'IsItNew':
                                   {'IsItNew':lambda x: ~(~x).any(),
                                    'ControlDate':'idxmin'}})
                        .droplevel(0, axis=1)
                        .reset_index()

df_['ControlDate'] = df.loc[df_['ControlDate'].values, 'ControlDate'].reset_index(drop=True)

  xName  IsItNew ControlDate
0  Car1    False  15/03/2018
1  Car2    False  25/05/2018
2  Car3     True  30/04/2018