数据转换-根据 aws 服务和 Pandas 中的百分比计算总成本

Question

I have dataframe like this: df1:

Product    line_item_product_code       Account    percentage
COMMON      AWSCloudTrail                AU-LOG     20  
COMMON      AWSGlue                      AU-LOG     30
COMMON      AWSQueueService              AU-LOG     50
COMMON      AWSSecretsManager            AU-PRD     40
COMMON      AmazonDynamoDB               AU-PRD     60

Second dataframe: df2

Account             Product               cost
AU-LOG       COMMON-PROD1                  10
AU-LOG       COMMON-PROD2                  12
AU-PRD       COMMON-PROD1                  14
AU-PRD       COMMON-PROD2                  16

Here total cost in df1 for a given account will match the total cost for that account in df2. The dataframe I want is:

So the calculation is to split total cost for a given account and product in df2 across various aws services based on percentage column in df1 and no of aws services are being used by a particular product.

Ex:In AU-LOG account there 3 different line_item_product_code and percentage is 20,30,50 .In df2 for AU-LOG account COMMON-PROD1 cost is $10 So this $10 will be splitted across 3 different line_item_product_code using the percentage mentioned in df1

Product          line_item_product_code  cost      Account    percentage 
COMMON-PROD1     AWSCloudTrail           2         AU-LOG       20  
COMMON-PROD1     AWSGlue                 3         AU-LOG       30
COMMON-PROD1     AWSQueueService         5         AU-LOG       50
COMMON-PROD2     AWSCloudTrail           2.4       AU-LOG       20  
COMMON-PROD2     AWSGlue                 3.6       AU-LOG       30
COMMON-PROD2     AWSQueueService         6         AU-LOG       50
COMMON-PROD1     AWSSecretsManager       5.6       AU-PRD       40
COMMON-PROD1     AmazonDynamoDB          8.4       AU-PRD       60
COMMON-PROD2     AWSSecretsManager       6.4       AU-PRD       40
COMMON-PROD2     AmazonDynamoDB          9.6       AU-PRD       60

How can I achieve this using pandas?

Answer 1

I think a simple merge should do, the multiply the cost column (which at this point contains the total product cost from df2) by the percentage / 100:

>>> df = pd.merge(df1.drop(columns=['Product']), df2, on='Account', how='left')
>>> df['cost'] = df['cost'] * df['percentage'] / 100
>>> df
  line_item_product_code Account  percentage       Product  cost
0          AWSCloudTrail  AU-LOG          20  COMMON-PROD1   2.0
1          AWSCloudTrail  AU-LOG          20  COMMON-PROD2   2.4
2                AWSGlue  AU-LOG          30  COMMON-PROD1   3.0
3                AWSGlue  AU-LOG          30  COMMON-PROD2   3.6
4        AWSQueueService  AU-LOG          50  COMMON-PROD1   5.0
5        AWSQueueService  AU-LOG          50  COMMON-PROD2   6.0
6      AWSSecretsManager  AU-PRD          40  COMMON-PROD1   5.6
7      AWSSecretsManager  AU-PRD          40  COMMON-PROD2   6.4
8         AmazonDynamoDB  AU-PRD          60  COMMON-PROD1   8.4
9         AmazonDynamoDB  AU-PRD          60  COMMON-PROD2   9.6