Data Manipulation in multiple columns in pandas dataframe

user_16702275 0
Viewed 190

I need to make a function, which takes input as dataframe, and dictionary{"Col_2":% change,"Col_3":% change} and it should make the changes to the dataframe.

I Have data frame like this

Date col_1 col_2 col_3
01/01/2022 90 100 300
01/02/2022 80 110 310
01/03/2022 92 120 400
01/04/2022 96 130 300
01/05/2022 99 150 500
01/06/2022 105 155 600

Now I pass the dictionary say,

{"col_2":-10,"col_3":10}

Then my desired out would look like this

Date col_1 col_2 col_3
01/01/2022 90 90 330
01/02/2022 80 99 341
01/03/2022 92 108 440
01/04/2022 96 113 330
01/05/2022 99 135 550
01/06/2022 105 139.5 660

I followed this code(but didn't get my desired result)

def per_change(df,col,per):
    m = 1+per/100
    df['col_2'] *= m
    return df
pandas dataframe dictionary python
3 Answers

You can specify columns names for processing in list and first divide values by 100, multiple by Series from dictionary and add to original DataFrame by DataFrame.add:

def per_change(df,cols,d):
    df[cols] = df[cols].add(df[cols].div(100).mul(pd.Series(d)), fill_value=0)
    return df

d =  {"col_2":-10,"col_3":10}
print (per_change(df,['col_1','col_2','col_3'],d))
         Date  col_1  col_2  col_3
0  01/01/2022   90.0   90.0  330.0
1  01/02/2022   80.0   99.0  341.0
2  01/03/2022   92.0  108.0  440.0
3  01/04/2022   96.0  117.0  330.0
4  01/05/2022   99.0  135.0  550.0
5  01/06/2022  105.0  139.5  660.0
user_29010020

You can use a Series of the dictionary with mul. By default the missing columns will be NaN, and you fill in the original data with combine_first.

The advantage of this solution is that you do not need to care about the existing columns, if will only multiply the ones present in the dictionary.

Numeric/string columns only:

(df+df.mul(pd.Series(d)/100)).combine_first(df)

Or to handle dataframes with non-numeric types:

(df+df.select_dtypes(include='number').mul(pd.Series(d)/100)).combine_first(df)

output:

         Date  col_1  col_2  col_3
0  01/01/2022   90.0   90.0  330.0
1  01/02/2022   80.0   99.0  341.0
2  01/03/2022   92.0  108.0  440.0
3  01/04/2022   96.0  117.0  330.0
4  01/05/2022   99.0  135.0  550.0
5  01/06/2022  105.0  139.5  660.0

input:

df = pd.DataFrame({'Date': ['01/01/2022', '01/02/2022', '01/03/2022', '01/04/2022', '01/05/2022', '01/06/2022'], 'col_1': [90, 80, 92, 96, 99, 105],
                   'col_2': [100, 110, 120, 130, 150, 155],
                   'col_3': [300, 310, 400, 300, 500, 600]})

d = {"col_2":-10,"col_3":10}
user_163434640

Function:

def changer(df, change_dic):
   new_df=df.copy()
   for col in change_dic.keys():
      new_df[col] += change_dic[col]*new_df[col]
   return new_df

and your edited version:

def per_change(df,col,per):
   m = 1+per/100
   df[col] *= m
   return df
user_72816750
Related