CODEWITHSUNDEEP
pythonpandasnumpy
Lynn
Lynn

Reputation: 4398

Multiple Aggregations on columns using Pandas

I have a dataset where I would like to display the percent as a result from grouping by 'type' as well as calculate the sum of two specific columns individually.

Data

free    use type    date
10      5   A1      1/1/2021
20      1   A1      1/1/2021
40      2   B2      1/1/2021
60      3   B2      1/1/2021

Desired

type    percent    use   free  date

A1      50         6     30    1/1/2021
B2      50         5     100   1/1/2021

Doing

percent = count/total *100
df.groupby(['type, 'date']).sum()

This gives me all but the percent. I am still researching. Any suggestion is appreciated

Upvotes: 1

Views: 50

Answers (2)

Amit Pathak
Amit Pathak

Reputation: 1377

You can try this approach -

Join the following pandas series object

percent_series = pd.DataFrame(df.type.value_counts())['type']/len(df)*100

Output

B     50.0
A1    50.0
Name: type, dtype: float64

Upvotes: 1

jezrael
jezrael

Reputation: 862761

Use Series.value_counts with normalize for percentage of original DataFrame and then use Series.map for aggregate values:

df1 = df.groupby(['type', 'date'], as_index=False).sum()

df1['perc'] =  df1['type'].map(df['type'].value_counts(normalize=True)).mul(100)
print (df1)
  type      date  free  use  perc
0   A1  1/1/2021    30    6  50.0
1   B2  1/1/2021   100    5  50.0

Upvotes: 2

Related Questions