Renaming Columns in Pandas DataFrames

This article explains how to rename a single or multiple columns in a Pandas DataFrame.

There are multiple different ways to rename columns and you’ll often want to perform this operation, so listen up.

Simple example

Create a Pandas DataFrame and print the contents.

df = pd.DataFrame({"num": [1, 2], "let": ["a", "b"]})
   num let
0    1   a
1    2   b

Let’s rename the columns to be number and letter so they’re more descriptive.

df.rename({"num": "number", "let": "letter"}, axis="columns", inplace=True)
   number letter
0       1      a
1       2      b

inplace=True causes the DataFrame to be mutated.

Rename single column

Create a DataFrame with first_name and last-name columns.

df = pd.DataFrame({"first_name": ["li", "karol"], "last-name": ["Fung", "G"]})
  first_name last-name
0         li      Fung
1      karol         G

We don’t want one column in our DataFrame to use underscores and the other to use hyphens. Rename the last-name column to be last_name.

df.rename({"last-name": "last_name"}, axis="columns", inplace=True)
  first_name last_name
0         li      Fung
1      karol         G

It’s easy to rename a single column in a DataFrame and leave the other column names unchanged.

Apply function to all column names

You can also apply a function to all column names. This is especially useful when you have a lot of columns or want to build a reusable transformation that can be applied to different DataFrames.

Create a DataFrame with spaces in the column names.

df = pd.DataFrame({"some place": ["hawaii", "costa rica"], "fun activity": ["surfing", "zip lining"]})
   some place fun activity
0      hawaii      surfing
1  costa rica   zip lining

Write a function that’ll replace all the spaces with underscores in the column names.

df.rename(lambda x: x.replace(" ", "_"), axis="columns", inplace=True)
   some_place fun_activity
0      hawaii      surfing
1  costa rica   zip lining

Next steps

Stick to the column renaming methods mentioned in this post and don’t use the techniques that were popular in earlier versions of Pandas.

This article intentionally omits legacy approaches that shouldn’t be used anymore.

The Pandas API is flexible and supports all common column renaming use cases:

  • renaming multiple columns with user specifed names
  • renaming some columns
  • applying a function to all column names

Be happy that Pandas is providing you with such a nice user interface!


Comments are closed, but trackbacks and pingbacks are open.