I have a confusion on the meaning of calculating sum in Pandas Pivot Table-CodePudding

pd.pivot_table(df1,index=['Staff'],columns=['Products'],values=['Price'],margins=True)

Results

Price
Products    A                 B              C              All
Staff                   
Staff_A 14472.000000    15877.777778    14982.352941    14890.196078
Staff_B 14775.000000    15620.000000    16330.000000    15815.789474
Staff_C 15293.333333    14262.500000    15214.285714    15000.000000
All     14779.545455    15231.818182    15426.470588    15099.000000

The result can be shown successfully however I do not understand how the values from the column and row All come from so can anybody explain how the value of sum comes from because I thought it should be the sum of the first three columns data added together?

CodePudding user response：

To illustrate:

import pandas as pd
import numpy as np

Create dummy dataframe:

df = pd.DataFrame({
  'id': ['a', 'b', 'c', 'a', 'b', 'c'],
  'stat': ['col1', 'col1', 'col1', 'col2', 'col2', 'col2'],
  'val': [2, 4, 6, 6, 8, 10],
})

This will give us:

>>> df
    id  stat    val
0   a   col1    2
1   b   col1    4
2   c   col1    6
3   a   col2    6
4   b   col2    8
5   c   col2    10

Pivoting without giving aggfunc argument

table = pd.pivot_table(df, index='id', columns='stat', values='val', margins=True)

The above will give us:

>>> table
stat    col1    col2    All
id          
a       2       6       4.0
b       4       8       6.0
c       6       10      8.0
All     4       8       6.0

Notice that column All and row All both giving us the mean of the columns and rows, respectively.

Pivoting and using aggfunc=np.sum:

new_table = pd.pivot_table(df, index='id', columns='stat', values='val', margins=True, aggfunc=np.sum)

This will give us:

>>> new_table
stat    col1    col2    All
id          
a       2       6       8
b       4       8       12
c       6       10      16
All     12      24      36