CODEWITHSUNDEEP
pythonpandasdataframe
Kaihua Hou
Kaihua Hou

Reputation: 368

How to only keep the first nth occurrence of a value in pandas data frame

I have a data frame as follows:

df = pd.DataFrame()
df['Name'] = ['Ankita', 'Ankita', 'Ankita', 'Ankita', 'Ankita', 'Yashvardhan', 'Yashvardhan', 'Yashvardhan', 'Yashvardhan', 'Yashvardhan']
df['Date'] = ['2014-10-07', '2015-03-30', '2015-12-07', '2015-12-09', '2017-01-30', '2017-01-30', '2018-02-19', '2018-02-23', '2018-11-19', '2020-01-23']
df['Value'] = [2200, 75, 100, 22, 98, 0.36, 57, 29, 1026, 1296]
df['Date'] = pd.to_datetime(df['Date'])

    Name    Date    Value
0   Ankita  2014-10-07  2200.00
1   Ankita  2015-03-30  75.00
2   Ankita  2015-12-07  100.00
3   Ankita  2015-12-09  22.00
4   Ankita  2017-01-30  98.00
5   Yashvardhan 2017-01-30  0.36
6   Yashvardhan 2018-02-19  57.00
7   Yashvardhan 2018-02-23  29.00
8   Yashvardhan 2018-11-19  1026.00
9   Yashvardhan 2020-01-23  1296.00

How can I only keep the earliest 3 rows of each unique name? i.e. how can I have the data frame end up like this:

    Name    Date    Value
0   Ankita  2014-10-07  2200.00
1   Ankita  2015-03-30  75.00
2   Ankita  2015-12-07  100.00
5   Yashvardhan 2017-01-30  0.36
6   Yashvardhan 2018-02-19  57.00
7   Yashvardhan 2018-02-23  29.00

And how can I only keep the latest two rows of each unique name? i.e. how can I have the data frame end up like this:

    Name    Date    Value
3   Ankita  2015-12-09  22.00
4   Ankita  2017-01-30  98.00
8   Yashvardhan 2018-11-19  1026.00
9   Yashvardhan 2020-01-23  1296.00

Thanks in advance!

Upvotes: 0

Views: 399

Answers (1)

SeaBean
SeaBean

Reputation: 23217

You can use .groupby() + GroupBy.head() and GroupBy.tail(), as follows:

df.groupby('Name').head(3)

          Name       Date    Value
0       Ankita 2014-10-07  2200.00
1       Ankita 2015-03-30    75.00
2       Ankita 2015-12-07   100.00
5  Yashvardhan 2017-01-30     0.36
6  Yashvardhan 2018-02-19    57.00
7  Yashvardhan 2018-02-23    29.00

df.groupby('Name').tail(2)

          Name       Date   Value
3       Ankita 2015-12-09    22.0
4       Ankita 2017-01-30    98.0
8  Yashvardhan 2018-11-19  1026.0
9  Yashvardhan 2020-01-23  1296.0

Upvotes: 3

Related Questions