How can i multiply a cell value of a dataframe based on two condition?-CodePudding

I have this dataframe

import numpy as np
import pandas as pd

data = {'month': ['5','5','6', '7'], 'condition': ["yes","no","yes","no"], 'amount': [500,200, 6000, 1873]}

and two values:

inflation5 = 1.05
inflation6 = 1.08

I need to know how can i multiply the cells of column 'amount' by the value inflation5 when the column 'month' value is 5 and the column 'condition' value is "yes", and also multiply the cells of column 'amount' by the value inflation6 when the column 'month' value is 6 and the column 'condition' value is "yes". Thanks

CodePudding user response：

Use numpy.select:

In [216]: df = pd.DataFrame(data)

In [218]: conds = [(df['condition'].eq('yes') & df['month'].eq('5')), (df['condition'].eq('yes') & df['month'].eq('6'))]

In [219]: inflation5 = 1.05
     ...: inflation6 = 1.08

In [220]: choices = [df['amount'].mul(inflation5), df['amount'].mul(inflation6)]

In [223]: df['amount'] = np.select(conds, choices, default=df['amount'])

In [224]: df
Out[224]: 
  month condition  amount
0     5       yes   525.0
1     5        no   200.0
2     6       yes  6480.0
3     7        no  1873.0

CodePudding user response：

For this I would run through with an np.where, should make it easily readable, and expandable especially if you wanted to change the condition with a function.

df = pd.DataFrame(data)
df['Inflation'] = np.where((df['month'] == '5') & (df['condition'] == 'yes'), inflation5, 1)
df['Inflation'] = np.where((df['month'] == '6') & (df['condition'] == 'yes'), inflation6, df['Inflation'])
df['Total_Amount'] = df['amount'].values * df['Inflation'].values

CodePudding user response：

I would suggest to use a different approach for efficiency.

Use a dictionary to store the inflations, then you can simply update in a single vectorial call:

inflations = {'5': 1.05, '6': 1.08}

mask = df['condition'].eq('yes')
df.loc[mask, 'amount'] *= df.loc[mask, 'month'].map(inflations)

NB. if you possibly have missing months in the dictionary, use df.loc[mask, 'month'].map(inflations).fillna(1) in place of df.loc[mask, 'month'].map(inflations)

output:

  month condition  amount
0     5       yes     525
1     5        no     200
2     6       yes    6480
3     7        no    1873