Convert dataframe from long to wide while retaining order of values (dtype datetime)

Question

I have a dataframe called df that looks similar to this (except the number of entries in the Visit Date column associated with each Client ID goes up to 74, and there are several hundred unique Client IDs - I have simplified it here).

Visit Date  Client ID
2016-05-25  C1009404
2016-06-30  C1009404
2016-07-14  C1009404
2016-07-20  C1009405
2016-08-03  C1009405
2016-08-08  C1009405
2016-08-10  C1009405
2016-08-15  C1009406
2016-08-17  C1009406
2016-08-24  C1009406

I want to convert it from long-to-wide, such that it looks like this:

Client ID  Visit_1     Visit_2     Visit_3     Visit_4
C1009404   2016-05-25  2016-06-30  2016-07-14
C1009405   2016-07-20  2016-08-03  2016-08-08  2016-08-10
C1009406   2016-08-15  2016-08-17  2016-08-24

I have tried the following code:

df_wide = df.groupby(['Client ID'], as_index=False).agg(lambda x: ', '.join(set(x.astype(str))))
df_wide = pd.concat([df_wide[['Client ID','ENROLLED_DT']], df_wide['VISIT_DT'].str.split(',', expand=True)], axis=1)
df_wide = df_wide.rename(columns={0: 'Visit_1', 1: 'Visit_2', 2: 'Visit_3', 3: 'Visit_4'})

It produces the desired result, but the dates are no longer in order. How do I do this but keep the dates in order, ascending from left to right?

Answer 1

You may need create another key for helping the pivot

df.assign(key=df.groupby('ClientID').cumcount()+1).\
    pivot('ClientID','key','VisitDate').\
      fillna('').\
        add_prefix('Visit_')
Out[152]: 
key           Visit_1     Visit_2     Visit_3     Visit_4
ClientID                                                 
C10094042  2016-05-25  2016-06-30  2016-07-14            
C10094056  2016-07-20  2016-08-03  2016-08-08  2016-08-10
C10094061  2016-08-15  2016-08-17  2016-08-24