CODEWITHSUNDEEP
pythonpandas
Dinosaurius
Dinosaurius

Reputation: 8628

How to group rows by column name and sort rows by datetime

I need to order rows by date and time grouping them by ID:

df =

ID   TYPE  DATETIME
1    A     12/18/2015 23:02:00
1    A     12/17/2015 12:00:09
1    A     12/18/2015 18:04:08
2    B     12/20/2015 15:03:05
2    B     12/18/2015 14:02:07

The result should be this one:

result =

ID   TYPE  DATETIME
1    A     12/17/2015 12:00:09
1    A     12/18/2015 18:04:08
1    A     12/18/2015 23:02:00
2    B     12/18/2015 14:02:07
2    B     12/20/2015 15:03:05

How can I do it?

df.groupby("ID").sort(...)

Upvotes: 1

Views: 60

Answers (2)

jezrael
jezrael

Reputation: 862611

You can use sort_values and specify columns by parameter by, ascending can be omit if need default ascending sorting:

#if df['DATETIME'].dtype != datetime, convert first
df.DATETIME = pd.to_datetime(df.DATETIME)

df = df.sort_values(by=['ID','TYPE','DATETIME'])
print (df)

   ID TYPE            DATETIME
1   1    A 2015-12-17 12:00:09
2   1    A 2015-12-18 18:04:08
0   1    A 2015-12-18 23:02:00
4   2    B 2015-12-18 14:02:07
3   2    B 2015-12-20 15:03:05

Upvotes: 1

RichSmith
RichSmith

Reputation: 940

You could try:

df.sort_values(['ID', 'DATETIME'], ascending=[True, True])

(I think df.sort() is now deprecated)

Upvotes: 1

Related Questions