CODEWITHSUNDEEP
pythonpandas
CandleWax
CandleWax

Reputation: 2219

How to sort only certain rows within a group in pandas?

I have a dataframe df that is grouped by ID. For each group, there is one row that has a flag that identifies that its the first instance 'First' == 1.

I ultimately want to sort each group by admit date, however, I need the row with 'First' == 1 to be the first row of that group regardless of its date. Then I want to sort the remaining rows based on admit date.

Sample df:

ID     admit     discharge    discharge_location   first
20     3-4-2018  3-6-2018     Home                 1
20     2-2-2018  2-6-2018     Home                 0
20     2-5-2018  2-23-2018    Home                 0
30     1-2-2018  2-3-2018     Home                 0
30     1-15-2018 1-18-2018    Home                 1
30     1-20-2018 1-24-2018    Home                 0

expected df:

ID     admit     discharge    discharge_location   first
20     3-4-2018  3-6-2018     Home                 1
20     2-2-2018  2-6-2018     Home                 0
20     2-5-2018  2-23-2018    Home                 0
30     1-15-2018 1-18-2018    Home                 1
30     1-2-2018  2-3-2018     Home                 0
30     1-20-2018 1-24-2018    Home                 0

My approach does not account for the 'first' column being first within the group.

df.sort_values(by=['ID','admit'], inplace=True)

This has been stumping me all day.

Upvotes: 2

Views: 1107

Answers (1)

ℕʘʘḆḽḘ
ℕʘʘḆḽḘ

Reputation: 19395

here it is bro,

df.sort_values(by=['ID', 'first', 'admit'], 
               ascending = [True, False, True],
               inplace = True)

Upvotes: 5

Related Questions