I wish to group and to convert datetimes to quarters within my dataset using Python

user_5942100 0
Viewed 29

I wish to group and to convert datetimes to quarters within my dataset, df.

Data

location    start       end         
aa/bb       1/1/2022    2/1/2022    
aa/bb       2/1/2022    3/1/2022    
aa/bb       3/1/2022    4/1/2022    
cc/dd       4/1/2022    5/1/2022    
cc/dd       5/1/2022    6/1/2022
cc/dd       6/1/2022    7/1/2022

Desired

location    start   end   
aa/bb       Q1 22   Q1 22   
cc/dd       Q2 22   Q2 22

Doing

df['start'] = pd.to_datetime(df['start'])
df['end'] = pd.to_datetime(df['end'])

df['start'] = pd.PeriodIndex(df['start'], freq='Q').strftime('Q%q.%y')
df['end'] = pd.PeriodIndex(df['end'], freq='Q').strftime('Q%q.%y')
pandas numpy python
2 Answers

Assuming all the dates of the same location belong to the same quarter

df['start'] = pd.to_datetime(df['start'])
df['end'] = pd.to_datetime(df['end'])

res = df.groupby('location').agg(lambda g: g.iloc[0].to_period('Q').strftime('Q%q.%y'))

A better way probably is

df['start'] = pd.PeriodIndex(df['start'], freq='Q').strftime('Q%q.%y')
df['end'] = pd.PeriodIndex(df['end'], freq='Q').strftime('Q%q.%y')

res = df.groupby('location').first()

Output

Input and output for both cases

>>> df

  location      start        end
0    aa/bb 2022-01-01 2022-02-01
1    aa/bb 2022-02-01 2022-03-01
2    aa/bb 2022-03-01 2022-04-01
3    cc/dd 2022-04-01 2022-05-01
4    cc/dd 2022-05-01 2022-06-01
5    cc/dd 2022-06-01 2022-07-01

>>> res

          start    end
location              
aa/bb     Q1.22  Q1.22
cc/dd     Q2.22  Q2.22
user_171206920

Try, coerce to datetime, extract period and format using dt.strftime. Then drop duplicates

   df =df.assign(start=pd.to_datetime(df['start']).dt.to_period('Q').dt.strftime('Q%q %y'), end= pd.to_datetime(df['end']).dt.to_period('Q').dt.strftime('Q%q %y')).drop_duplicates()
   



  location  start    end
0    aa/bb  Q1 22  Q1 22
2    aa/bb  Q1 22  Q2 22
3    cc/dd  Q2 22  Q2 22
5    cc/dd  Q2 22  Q3 22
user_89869750
Related