Calculated Columns in Pandas

Create new columns in pandas

Pandas has got two very useful functions called groupby and transform. In this TIL, I will demonstrate how to create new columns from existing columns.

First of all, I create a new data frame here.

df = pd.DataFrame( {'city':['London','London','Berlin','Berlin'], 'rent': [1000, 1400, 800, 1000]} )

which looks like

	city	rent
0	London	1000
1	London	1400
2	Berlin	800
3	Berlin	1000

I will create a new column called total, which will host the total rents of the corresponding cities.

df['total'] = df.groupby('city').transform('sum')

	city	rent	total
0	London	1000	2400
1	London	1400	2400
2	Berlin	800	1800
3	Berlin	1000	1800

Next, I will create a new column called percent which will contain the percentage

df['percent'] = df['rent']/df['total']

	city	rent	total	percent
0	London	1000	2400	0.416667
1	London	1400	2400	0.583333
2	Berlin	800	1800	0.444444
3	Berlin	1000	1800	0.555556

Planted: 2018-05-20 by Lei Ma;

Dynamic Backlinks to til/programming/pandas/pandas-new-column-from-other:

Deal with Rare Categories Using Pandas

Deal with rare categories using pandas

L Ma (2018). 'Calculated Columns in Pandas', Datumorphism, 05 April. Available at: https://datumorphism.leima.is/til/programming/pandas/pandas-new-column-from-other/.