CODEWITHSUNDEEP
pythonpandas
biwia
biwia

Reputation: 397

pandas: create new column group by others

i want to create new column that will calculate a ratio groupby date and count like shown below in the result of my new column new.

   date    count    place       moy       new
2021-07-04  396       NY        30.     30/(30+10+20+40)
2021-07-04  396       OX        10.     10/(30+10+20+40)
2021-07-04  396       CA        20      20/(30+10+20+40)
2021-07-04  396       LA        40      40/(30+10+20+40)
2021-07-05  592       NY        30.     30/(30+50+10+30)
2021-07-05  592       OX        50.     50/120
2021-07-05  592       NY        10.     10/120
2021-07-05  592       OX        30.     30/12O

Upvotes: 0

Views: 56

Answers (1)

Anurag Dabas
Anurag Dabas

Reputation: 24324

Try via groupby() and transform():

df['new']=df['moy']/df.groupby(['date','count'])['moy'].transform('sum')
#OR
df['new']=df['moy'].div(df.groupby(['date','count'])['moy'].transform('sum'))

If needed you can round the values by using round() method:

df['new']=df['new'].round(2)

Now If you print df you will get your desired output

Upvotes: 1

Related Questions