Get Maximum in each Group – Pandas Groupby

You can use Pandas groupby to group the underlying data on one or more columns and estimate useful statistics like count, mean, median, min, max etc. In this tutorial, we will look at how to get the maximum value for each group in pandas groupby with the help of some examples.

Pandas Groupby Maximum

To get the maximum value of each group, you can directly apply the pandas max() function to the selected column(s) from the result of pandas groupby. The following is a step-by-step guide of what you need to do.

1. Group the dataframe on the column(s) you want.
2. Select the field(s) for which you want to estimate the maximum.
3. Apply the pandas max() function directly or pass ‘max’ to the agg() function.

The following is the syntax –

# groupby columns on Col1 and estimate the maximum value of column Col2 for each group
df.groupby([Col1])[Col2].max()
# alternatively, you can pass 'max' to the agg() function
df.groupby([Col1])[Col2].agg('max')

Examples

Let’s look at some examples of using the above syntax to get the max value for each group in pandas. First, we will create a sample dataframe that we will be using throughout this tutorial.

import pandas as pd

# create a dataframe of car models by two companies
df = pd.DataFrame({
    'Company': ['A', 'A', 'A', 'B', 'B', 'B', 'B'],
    'Model': ['A1', 'A2', 'A3', 'B1', 'B2', 'B3', 'B4'],
    'Year': [2019, 2020, 2021, 2018, 2019, 2020, 2021],
    'Transmission': ['Manual', 'Automatic', 'Automatic', 'Manual', 'Automatic', 'Automatic', 'Manual'],
    'EngineSize': [1.4, 2.0, 1.4, 1.5, 2.0, 1.5, 1.5],
    'MPG': [55.4, 67.3, 58.9, 52.3, 64.2, 68.9, 83.1]
})
# display the dataframe
df

Output:

Now we have a dataframe containing the specifications of the car models by two different companies. The “EngineSize” column is the size of the engine in litres and the “MPG” column is the mileage of the car in miles-per-gallon.

1. Groupby Maximum of a single column

Let’s get the maximum value of mileage “MPG” for each “Company” in the dataframe df. To do this, group the dataframe on the column “Company”, select the “MPG” column, and then apply the max() function.

# max MPG for each Company
df.groupby('Company')['MPG'].max()

Output:

Company
A    67.3
B    83.1
Name: MPG, dtype: float64

You can see that maximum mileage “MPG” for company “B” is higher than that of company “A”.

Alternatively, you can also use the pandas agg() function on the resulting groups.

# max MPG for each Company
df.groupby('Company')['MPG'].agg('max')

Output:

Company
A    67.3
B    83.1
Name: MPG, dtype: float64

We get the same result as above.

You can also group the data on multiple columns (to get more granular groups) and then compute the max for each group. For example, let’s group the data on “Company” and “Transmission” and get the maximum “MPG” for each group.

# max MPG for each Company at a transmission level
df.groupby(['Company', 'Transmission'])['MPG'].max()

Output:

Company  Transmission
A        Automatic       67.3
         Manual          55.4
B        Automatic       68.9
         Manual          83.1
Name: MPG, dtype: float64

2. Groupby Maximum of multiple columns

You can also get the maximum value of different columns for each group resulting from pandas groupby. For example, let’s get the maximum of mileage “MPG” and “EngineSize” for each “Company” in the dataframe df.

# max MPG and EngineSize for each Company
df.groupby('Company')[['MPG', 'EngineSize']].max()

Output:

Here we selected the columns that we wanted to compute the maximum on from the resulting groupby object and then applied the max() function. We already know that the maximum “MPG” is higher for company “B” but the maximum “EngineSize” is equal for both the companies.

Let’s now do the same thing using the pandas agg() function.

# max MPG and EngineSize for each Company
df.groupby('Company')[['MPG', 'EngineSize']].agg('max')

Output:

We get the same result as above.

With this, we come to the end of this tutorial. The code examples and results presented in this tutorial have been implemented in a Jupyter Notebook with a python (version 3.8.3) kernel having pandas version 1.0.5