In this article, we will cover how to count NaN and non-NaN values in Pandas DataFrame or column.

Missing values in Pandas are represented by NaN - not a number but sometimes are referred as:

  • NA
  • None
  • null

We will see how to count all of them.

Here is how to count NaN and non NAN values in Pandas:

(1) Count NA Values in Pandas DataFrame

df.count()

(2) Count non NA Values in DataFrame

df.isna().sum()

(3) Count NA Values in Pandas column

df['col1'].count()

(4) Count non NA Values in DataFrame

df['col1'].isna().sum()

Count Na Values

To count the number of NaN values in a Pandas DataFrame or Series, we can

  • use the .isna() method
  • then sum the resulting Boolean values( 1 = True, 0 = False):

DataFrame

To count Na values in the whole Pandas DataFrame we can apply isna() on every column:

import pandas as pd

df = pd.DataFrame({'col1': ['a', None, 3, None, 5],
          		 'col2': [None, 7, 'b', 3, 4]})

na_count = df.isna().sum()
print(na_count)

result:

col1    2
col2    1
dtype: int64

Column

To count Na values in Pandas column we can sum Na values in the column:

df['col1'].isna().sum()

The result is the number of the Na values in this column - 2.

Count non Na Values

To count the number of non-NaN values in a Pandas DataFrame or Series, we can use methods:

  • pandas.DataFrame.count
  • pandas.Series.count

DataFrame

Method count return number of the non Na values for the whole DataFrame:

df.count()

result:

col1    3
col2    4
dtype: int64

Not that this will count the non Na values column wise. For row-wise refer to the next section.

Row-wise

We can count non Na values in a given Pandas DataFrame row-wise by using parameter axis=1 and pass it to count method:

df.count(axis=1)

result is non Na values in each row:

0    1
1    1
2    2
3    1
4    2
dtype: int64

Column

To count non NaN values in Pandas column we can use the Series count method:

df['col1'].count()

as output we get the number of non Na values in col1: 3.

The code above is equivalent to:

df['col1'].notna().sum()

Count non Na values - describe()

We can use Pandas method describe to count non Na values in the whole DataFrame or column by:

df['col1'].describe()

result:

count 	3
unique	3
top   	a
freq  	1
Name: col1, dtype: object

Count Na values - value_counts()

We can check the number of Na or non Na values also by using the method: value_counts(). To do so we need to pass parameter dropna=False:

df['col1'].value_counts(dropna=False)

result:

None	2
a   	1
3   	1
5   	1
Name: col1, dtype: int64

Count percent of missing values

To count the percent of the missing values in each column of Pandas DataFrame we can use:

  • isna()
  • chain method mean()
df.isna().mean()

This will give us the percent of the Na values in the selected columns:

col1	0.4
col2	0.2
dtype: float64

Multiply by 100 to get value between 0 and 100:

df.isna().mean() * 100

result:

col1	40.0
col2	20.0
dtype: float64

Count Na and non Na values in column

To count both Na and non Na values in Pandas column we can use isna in combination with value_coutns() method:

df['col1'].isna().value_counts()

The results is number of Na and non Na values in this column:

False	3
True 	2
Name: col1, dtype: int64

Conclusion

In this article we covered how to count the number of NaN and non NaN values in Pandas DataFrame. We saw how to count row and column-wise.

We count Na values for the whole DataFrame or a single column. Finally we saw how to calculate the percent of missing values and count Na / non Na values in a column.