CODEWITHSUNDEEP
pythonpandaspandas-groupby
regretoverflow
regretoverflow

Reputation: 2153

add group index value to groups

I have a dataframe ('df') something like (simplified for this example):

index |     timestamp    | value
================================
001   | 2020-09-20 07:00 | 1.4
002   | 2020-09-20 07:00 | 1.5
001   | 2020-09-20 09:00 | 1.6
002   | 2020-09-20 09:00 | 1.4
001   | 2020-09-20 11:00 | 1.23
002   | 2020-09-20 11:00 | 1.46

If I do this: grouped = df.groupby('timestamp') I now have a groupby with three groups. I need to now add a 'date_time_trigger' column that contains a value based on the index of the group:

index |     timestamp    | value | date_time_trigger
================================================
001   | 2020-09-20 07:00 | 1.4   | triggergroup1
002   | 2020-09-20 07:00 | 1.5   | triggergroup1
001   | 2020-09-20 09:00 | 1.6   | triggergroup2
002   | 2020-09-20 09:00 | 1.4   | triggergroup2
001   | 2020-09-20 11:00 | 1.23  | triggergroup3
002   | 2020-09-20 11:00 | 1.46  | triggergroup3

I then need to combine the groups back into the original dataframe. I've tried simply adding column to the original data frame, then changing its value inside a group iterator:

idx=0
df['date_time_trigger']='foo'
grouped = df.groupby('timestamp')
for name,group in grouped:
  idx=idx+1
  group['date_time_trigger']='triggergroup'+str(idx)

And as far as I can tell, the value of date_time_trigger is being set inside each group, but now I need to recombine the group into the original dataframe df to continue with my process. The only way I can find in the docs is to apply some type of aggregation, like mean or avg, but I just needed the groupby to add the labels to each group instance. How do I get my dataframe back?

Upvotes: 1

Views: 120

Answers (1)

BENY
BENY

Reputation: 323226

Check with factorize

df['group'] = df['timestamp'].factorize()[0]+1

Method two

df.groupby('timestamp').ngroup().add(1).astype(str).radd('triggergroup')

0    triggergroup1
1    triggergroup1
2    triggergroup2
3    triggergroup2
4    triggergroup3
5    triggergroup3
dtype: object

Upvotes: 1

Related Questions