How to Drop Rows with NaN Values in Pandas DataFrame – Data to Fish (2024)

To drop rows with NaN (null) values in Pandas DataFrame:

df.dropna()

To drop only the rows where all the values are NaN:

df.dropna(how="all")

Steps to Drop Rows with NaN Values in Pandas DataFrame

Step 1: Create a DataFrame with NaN Values

Create a DataFrame with NaN values:

import pandas as pdimport numpy as npdata = {"col_a": [1, 2, np.nan, 4], "col_b": [5, np.nan, np.nan, 8], "col_c": [9, 10, 11, 12] }df = pd.DataFrame(data)print(df)

As can be observed, the second and third rows now have NaN values:

 col_a col_b col_c0 1.0 5.0 91 2.0 NaN 102 NaN NaN 113 4.0 8.0 12

Step 2: Drop the Rows with the NaN Values in Pandas DataFrame

Use df.dropna() to drop all the rows with the NaN values in the DataFrame:

import pandas as pdimport numpy as npdata = {"col_a": [1, 2, np.nan, 4], "col_b": [5, np.nan, np.nan, 8], "col_c": [9, 10, 11, 12] }df = pd.DataFrame(data)df_dropped = df.dropna()print(df_dropped)

There results are two rows without any NaN values:

 col_a col_b col_c0 1.0 5.0 93 4.0 8.0 12

Noticed that those two rows no longer have a sequential index. It’s currently 0 and 3. You can then reset the index to start from 0 and increase sequentially.

Step 3 (Optional): Reset the Index

The general syntax to reset an index in Pandas DataFrame:

df.reset_index(drop=True)

The complete script to drop the rows with the NaN values, and then reset the index:

import pandas as pdimport numpy as npdata = {"col_a": [1, 2, np.nan, 4], "col_b": [5, np.nan, np.nan, 8], "col_c": [9, 10, 11, 12] }df = pd.DataFrame(data)df_dropped = df.dropna()df_reset = df_dropped.reset_index(drop=True)print(df_reset)

The index now starts from 0 and increases sequentially:

 col_a col_b col_c0 1.0 5.0 91 4.0 8.0 12

Drop Rows Where all the Values are NaN

Here is an example of a DataFrame where all the values are NaN for the third row:

import pandas as pdimport numpy as npdata = {"col_a": [1, 2, np.nan, 4], "col_b": [5, np.nan, np.nan, 8], "col_c": [9, 10, np.nan, 12] }df = pd.DataFrame(data)print(df)

As can be seen, all the values are NaN for the third row:

 col_a col_b col_c0 1.0 5.0 9.01 2.0 NaN 10.02 NaN NaN NaN3 4.0 8.0 12.0

Use df.dropna(how=”all”) to drop only the row/s where all the values are NaN:

import pandas as pdimport numpy as npdata = {"col_a": [1, 2, np.nan, 4], "col_b": [5, np.nan, np.nan, 8], "col_c": [9, 10, np.nan, 12] }df = pd.DataFrame(data)df_dropped = df.dropna(how="all")print(df_dropped)

The result:

 col_a col_b col_c0 1.0 5.0 9.01 2.0 NaN 10.03 4.0 8.0 12.0