How to create a list from a Pandas DataFrame?

In today’s Python data analysis tutorial we will how to make a list from Pandas objects.

We’ll look into several use cases:

  • Convert one DataFrame column to a list.
  • Get multiple column values into a list.
  • Create a list from an entire DataFrame.

Create Pandas example data

Below is a short Python snippet to create a simple DataFrame that you can use to follow along this example.

import pandas as pd
city =  ['Tokyo', 'Paris', 'Osaka','Sidney']
office =  ['West', 'West', 'South', 'East']
sales = [110,103,145,190]

offices = dict(city=city, office=office, revenue=sales)
revenue = pd.DataFrame (offices)


Here’s our example data:


Convert a Series to a list

To make a list out of a DataFrame column, we can simply use the Series to_list() method as shown in the snippet below.

city_list = revenue['city'].to_list()

And the result will be:

['Tokyo', 'Paris', 'Osaka', 'Sidney']

You can obviously validate the we indeed create a list by using the type function:


Make a list from multiple columns

In this case, we’ll first define a subset of columns that we’ll slice into a list; then we’ll call the DataFrame.values.tolist() method.

# multiple columns to list
cols = ['city', 'office']

Convert entire DataFrame to list

In a similar fashion we are able to get the entire DataFrame into our Python list. Code is shown below:

df_list = revenue.values.tolist()

The result is obviously a list of lists containing all the DataFrame rows. Each is exported as a list.

[['Tokyo', 'West', 110], ['Paris', 'West', 103], ['Osaka', 'South', 145], ['Sidney', 'East', 190]]

You can now go the other way around and create a DataFrame from the list of lists by running the following code:

revenue2 = pd.DataFrame(df_list, columns = ['city', 'office', 'revenue'])

DataFrame row to list

It is also easy to save a specific DataFrame row to a Pyhon list. The code below captures the first Dataframe row values into a list.


I can also make a list out of several DataFrame rows by slicing the DataFrame or using the loc or iloc indexers:


# or using iloc