In a pandas dataframe, how can I filter the rows based on a column value, do calculation and assign the result to a new column?

Question

For example, I have a dataframe of four people, divided into group A and B. Now I want to filter out the group B, divide their Point by half, and assign the result to a new column named as 'Point_new'.

I am using the codes like the following

import pandas as pd

data = {'Name':['Tom', 'Nick', 'Krish', 'Jack'],
        'Group':['A', 'B', 'B', 'A'],
        'Point':[20, 21, 19, 18]}

df = pd.DataFrame(data)

df['Point_new'] = ''

df[df['Group']=='B']['Point_new'] = df[df['Group']=='B']['Point'] / 2

From output of the above codes, the Point_new column is not filled with the calculation result. I wonder why is that, and how could I do it properly.

Answer 1

You can do it like this, the key is to use loc -

import pandas as pd

data = {'Name':['Tom', 'Nick', 'Krish', 'Jack'],
        'Group':['A', 'B', 'B', 'A'],
        'Point':[20, 21, 19, 18]}

df = pd.DataFrame(data)


df.loc[(df['Group'] == 'B'),'Point_new'] = df.loc[df['Group']=='B','Point'] / 2


print(df)

    Name Group  Point  Point_new
0    Tom     A     20        NaN
1   Nick     B     21       10.5
2  Krish     B     19        9.5
3   Jack     A     18        NaN

Answer 2

It's because you're trying to set a value on a slice of a DataFrame. See more info here: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

Try this:

df.loc[df['Group']=='B', 'Point_new'] = df[df['Group']=='B']['Point'] / 2

Which yields:

	Name	Group	Point	Point_new
0	Tom	A	20
1	Nick	B	21	10.5
2	Krish	B	19	9.5
3	Jack	A	18

Answer 3

You can try:

df = df.assign(Point_New = df[df.Group.eq('B')].Point.div(2))

---

    Name    Group   Point   Point_New
0   Tom     A       20      NaN
1   Nick    B       21      10.5
2   Krish   B       19      9.5
3   Jack    A       18      NaN