How to Get Rows or Columns with NaN (null) Values in Pandas DataFrame

Let’s learn how to get rows or columns with one or more NaN values in a Pandas DataFrame.

1. Get rows with NaN

You can use isna() or isnull() to get all rows with NaN values.

df[df.isna().any(axis=1)]
df[df.isnull().any(axis=1)]

isnull() is an alias of isna(). Many prefer isna() for semantic consistency (e.g. np.isnan(), dropna(), and fillna() all handle missing values).

2. Get columns with NaN

Let’s retrieve all rows of all columns that contain a NaN value.

df.loc[:, df.isna().any()]

3. Get each row’s NaN status

Given a single column, pd.isnull() will tell us the NaN status of each row (i.e. whether or not it contains any NaN values in that column).

pd.isnull(df['col'])

4. Get NaN count per column

You can also get the count of NaN values per column using isnull() and sum().

df.isnull().sum()