How to drop the last rows in Pandas DataFrames?

by

In this tutorial we’ll learn how to remove the one or several last rows of a DataFrame.

We’ll be touching on several cases:

  • Getting the last (or last n) rows in a DataFrame.
  • Removing the last (or last n) rows from the DataFrame.
  • Dropping all rows except the first row
  • Drop the last column

Example data

We’ll start by defining a simple DataFrame that you can use in order to follow along with this exercise.

import pandas as pd

month = ['March', 'March', 'March', 'April', 'April', 'March']
language = ['Java', 'Javascript', 'Javascript', 'R', 'R', 'Javascript']
salary = [138.0, 138.0, 108.0, 109.0, 109.0, 127.0]
salaries = dict(month=month, language=language, salary = salary)
salary_df = pd.DataFrame(data=salaries)
salary_df

Here’s our small DataFrame

monthlanguagesalary
0MarchJava138.0
1MarchJavascript138.0
2MarchJavascript108.0
3AprilR109.0
4AprilR109.0
5MarchJavascript127.0

Get the last row of a Pandas DataFrame

We are well familiar with the head() DataFrame method, that allows to fetch the first rows of a DataFrame. Conversely, we also have the tail() method, the allows to retrieve the last:

salary_df.tail(1)

Will retrieve the last row:

monthlanguagesalary
5MarchJavascript127.0

Note that we can retrieve more rows from the DataFrame tail. In this example – the last 3 rows.

n=3
salary_df.tail(n)

Drop the last row from the DataFrame

We can now use the drop() function to easily remove the last row from our DataFrame

last_row = salary_df.tail(1).index
salary_df.drop (last_row, inplace=True)

The inplace=True persist the changes we have done in the original DataFrame. If you are not interested in modifying your DataFrame, you can simply assign the change data into a new DataFrame:

new_df = salary_df.drop (last_row)

Drop the last n rows

In a similar fashion: 

n=3
last_n_rows = salary_df.tail(n).index
salary_df.drop (last_n_rows, inplace=True)

Removing all rows except the first

We can easily drop all DataFrame rows, but leave the first:

all_rows_except_first = salary_df.tail(len(salary_df)-1).index
salary_df.drop (all_rows_except_first)

Here’s our result:

monthlanguagesalary
0MarchJava138.0

Removing the last column off your DataFrame

So far, we dealt with rows, but using a similar technique we can also get rid of specific columns.

#find the last element in the column index
last_col = salary_df.columns[-1]

new_df = salary_df.drop(cols, axis=1)

Note the usage of axis=1, to determine that we are interested in removing a column and not a row index.

Additional learning