Pandas – Delete All Columns Except Some Columns

In this tutorial, we will look at how to delete all the columns of a pandas dataframe except one (or some) specific column(s) with the help of some examples.

How to delete all columns except one (or more) specific column(s)?

To delete all the columns except some specific ones from a pandas dataframe, you can filter the dataframe such that it contains only those specific columns.

The are multiple ways to filter a dataframe for specific columns in Pandas –

• You can use the square brackets notation to select the columns that you want. For example, if you want to retain only columns “Col1” and “Col2” from the dataframe df, use df[['Col1', 'Col2']].
• Using the pandas dataframe filter() function.
• Using .loc property to keep only the columns you want.

Examples

Let’s now look at some examples of using the above methods to remove all columns except some from a pandas dataframe.

For this tutorial, we will be using the Pokemon dataset to create our dataframe and show the usage of the different methods.

import pandas as pd

# load the pokemon dataset
df = pd.read_csv('data/Pokemon.csv')
# display the dataframe head
df.head()

Output:

Here, we loaded the Pokemon dataset as a dataframe using the read_csv() function in pandas. You can see that the dataframe has 13 columns with different information and attributes about a pokemon.

Example 1 – Remove all columns except some using square bracket notation

Let’s remove all the columns except the “Name” column from the above dataframe using the square bracket notation.

In this method, we use the list of columns we want to select (or retain) inside a pair of square brackets with the dataframe.

# keep only the "Name" column
df = df[["Name"]]
# display the dataframe head
df.head()

Output:

You can see that the resulting dataframe contains only the “Name” column.

You can similarly use this method to remove all the columns except some specific ones. For example, let’s reload the dataset and remove all the columns except “Name” and “Type 1”.

# load the pokemon dataset
df = pd.read_csv('data/Pokemon.csv')
# keep only the "Name" and "Type 1" columns
df = df[["Name", "Type 1"]]
# display the dataframe head
df.head()

Output:

The dataframe now has only the “Name” and “Type 1” columns.

Example 2 – Using the filter() method

The pandas dataframe filter() method is used to filter a dataframe on rows and/or columns.

Pass the list of column names you want to filter (or retain) as an argument to the filter() method (it filters columns by default).

Let’s take the same use case from above. We’ll load the entire pokemon dataset and then remove all the columns except “Name” and “Type 1”.

# load the pokemon dataset
df = pd.read_csv('data/Pokemon.csv')
# keep only the "Name" and "Type 1" columns
df = df.filter(["Name", "Type 1"])
# display the dataframe head
df.head()

Output:

We get the same result as above.

Example 3 – Retain columns using the .loc property

The loc dataframe property is used to slice (or filter) a dataframe based on its row and column labels.

Let’s take the same use case from the above examples. We’ll load the entire pokemon dataset and then remove all the columns except “Name” and “Type 1”.

# load the pokemon dataset
df = pd.read_csv('data/Pokemon.csv')
# keep only the "Name" and "Type 1" columns
df = df.loc[:, ["Name", "Type 1"]]
# display the dataframe head
df.head()

Output:

We get the same result as above.

Summary

In this tutorial, we looked at some methods to delete all the columns except some (one or more) columns from a pandas dataframe by filtering the dataframe. The following are the key takeaways –

• You can use the square bracket notation to select (or retain) the specific columns you want.
• Alternatively, you can use the pandas filter() function to filter the dataframe for specific columns.
• You can also use the dataframe .loc property to filter a dataframe using column names.

You might also be interested in –

• Pandas – Filter DataFrame for multiple conditions
• Pandas – Delete rows based on column values
• Filter DataFrame rows on a list of values
• Pandas – Get dataframe summary with info()
• Pandas – Change Column Names to Uppercase
• Pandas – Change Column Names to Lowercase