How to Wrap/Break Long Column Names in Pandas Dataframe   

To wrap or break long column names in Pandas we can use module textwrap and map the column names with new line symbols:

(1) Wrap DataFrame column names

import textwrap
cols_wrap = [textwrap.wrap(x, width=20) for x in df.columns]
cols_wrap = {' '.join(words) : '<br>'.join(words) for words in cols_wrap}
cols_wrap

(2) Truncate column names

import textwrap
cols_wrap = {x: textwrap.wrap(x, width=15)[1] for x in df.columns}

This will prevent formatting issues or horizontal overflow on displayed data. Let's see it in more details and examples:

Data

Let's use this data:

import pandas as pd

df = pd.DataFrame({
  "Test Data Type N0 extract 1": [0, 1, 2, 3],
  "Test Data Type N101 extract 1": [3, 5, 7, 9],
  "Prod Data Type N0 extract 0": [1, 2, 3, 4],
  "Prod Data Type N101 extract 0": [0.5, 1.0, 1.5, 2.0],
})

which will have long names. If you work with 20+ columns this might be visually hard to digest:

1. Wrap column names

We can wrap every column name no matter is it OK or too long, by inserting <br> to break them:

import textwrap
cols_wrap = [textwrap.wrap(x, width=20) for x in df.columns]
cols_wrap = {' '.join(words) : '<br>'.join(words) for words in cols_wrap}
cols_wrap

This will create a dictionary:

{'Test Data Type N0 extract 1': 'Test Data Type N0<br>extract 1',
 'Test Data Type N101 extract 1': 'Test Data Type N101<br>extract 1',
 'Prod Data Type N0 extract 0': 'Prod Data Type N0<br>extract 0',
 'Prod Data Type N101 extract 0': 'Prod Data Type N101<br>extract 0'}

df.rename(columns=cols_wrap).style.format()

Now we can display the DataFrame with wrapped column names:

• we can control the lenght of the wrap by - width=20
• shorter columns will remain the same
• the <br> works in Jupyterlab in combination with .style.format()
• original data is unchanged

2. Truncate Column names

We can also break the longer column names by similar approach:

import textwrap
cols_wrap = {x: textwrap.wrap(x, width=15)[1] for x in df.columns}
cols_wrap

this time we will have shorter names which consists only from the last part of the wrap:

{'Test Data Type N0 extract 1': 'N0 extract 1',
 'Test Data Type N101 extract 1': 'N101 extract 1',
 'Prod Data Type N0 extract 0': 'N0 extract 0',
 'Prod Data Type N101 extract 0': 'N101 extract 0'}

result:

3. Transpose for better vertical readability

If the dataset is small, sometimes transposing helps:

print(df.T.to_string())

or by printing:

print(df.T.to_string())

This prints the column names as row labels, which makes even long names easier to read:

Resource

• Tutorials on Pandas DataFrame columns
• Break/wrap long text of column names in Pandas dataframe plain text to_string output?