pandas-calculate percentage in specfic group

Question

My input dataframe: df1

line_item_product_code  cost Account     
AWSCloudTrail            10  AU-MGT   
AWSGlue                  12  AU-MGT  
AWSQueueService          15  AU-MGT   
AWSSecretsManager        20  AU-PRD  
AmazonDynamoDB           30  AU-PRD

I want to calculate the percentage based on specific account

Output dataframe:

line_item_product_code  cost Account     percentage
AWSCloudTrail            10  AU-MGT      27.02 
AWSGlue                  12  AU-MGT      32.43
AWSQueueService          15  AU-MGT      40.54
AWSSecretsManager        20  AU-PRD      40
AmazonDynamoDB           30  AU-PRD      60

How can I achieve this in pandas? Example Calculation for AU-MGT and AWSCloudTrail:

percentage = 10*100/(10+12+15)

Does this answer your question? [Pandas percentage of total with groupby](https://stackoverflow.com/questions/23377108/pandas-percentage-of-total-with-groupby) — mozway, Jul 07 '21 at 17:17

Cimbali · Accepted Answer · 2021-07-07T17:02:58.673

3

A simple groupby + transform should do the trick:

>>> df['percentage'] = df['cost'] / df.groupby('Account')['cost'].transform('sum') * 100
>>> df
  line_item_product_code  cost Account  percentage
0          AWSCloudTrail    10  AU-MGT   27.027027
1                AWSGlue    12  AU-MGT   32.432432
2        AWSQueueService    15  AU-MGT   40.540541
3      AWSSecretsManager    20  AU-PRD   40.000000
4         AmazonDynamoDB    30  AU-PRD   60.000000

edited Jul 07 '21 at 17:02

answered Jul 07 '21 at 16:58

Cimbali

11,012
1
39
68

[This related question](https://stackoverflow.com/questions/23377108/pandas-percentage-of-total-with-groupby) addresses the efficiency of different methods to calculate percentages per group. – mozway Jul 07 '21 at 17:17

pandas-calculate percentage in specfic group

1 Answers1