notna(), aliased identically as notnull(), is functionally equivalent to negating the result of isna(), provided as a separate, readable method so code that specifically cares about present values doesn't need an extra negation to express that intent. It's commonly used with boolean indexing to select only the rows that do have a value in a specific column, the mirror image of using isna() to find the rows that don't.

1Understanding df.notna()

💡

df.notna() and negating df.isna() are exactly equivalent — use notna() directly whenever your logic is naturally phrased as selecting the values that are present, instead of an extra negation on isna().

editor.html

import pandas as pd
import numpy as np

df = pd.DataFrame({"email": ["a@x.com", np.nan, "c@x.com"]})
print(df["email"].notna())

localhost:3000

2Practical Example

Here is a real-world application of df.notna() showing how it is used in production Pandas code.

editor.html

import pandas as pd
import numpy as np

df = pd.DataFrame({"name": ["Alice", "Bob", "Carol"], "email": ["a@x.com", np.nan, "c@x.com"]})
has_email = df[df["email"].notna()]
print(has_email)

localhost:3000

3Best Practices

Follow these guidelines when working with df.notna():

1. Use notna() directly when your logic is naturally about selecting present values, rather than writing an explicit negation of isna()

2. Combine notna() with boolean indexing to select only rows with a value in a specific required column

3. Use notna().sum() to count non-missing values per column, the mirror image of isna().sum() for counting missing ones

⚠️

Tip: df.notna() and negating df.isna() are exactly equivalent — use notna() directly whenever your logic is naturally phrased as selecting the values that are present, instead of an extra negation on isna().

editor.html

import pandas as pd
import numpy as np

df = pd.DataFrame({"email": ["a@x.com", np.nan, "c@x.com"]})
print(df["email"].notna())

localhost:3000

Examples

Example 01Basic Usage

import pandas as pd
import numpy as np

df = pd.DataFrame({"email": ["a@x.com", np.nan, "c@x.com"]})
print(df["email"].notna())

Example 02Advanced Example

import pandas as pd
import numpy as np

df = pd.DataFrame({"name": ["Alice", "Bob", "Carol"], "email": ["a@x.com", np.nan, "c@x.com"]})
has_email = df[df["email"].notna()]
print(has_email)

Best Practices

Use notna() directly when your logic is naturally about selecting present values, rather than writing an explicit negation of isna()
Combine notna() with boolean indexing to select only rows with a value in a specific required column
Use notna().sum() to count non-missing values per column, the mirror image of isna().sum() for counting missing ones

Interview Question

Is there any actual behavioral difference between df.notna() and manually negating the result of df.isna()?▼

Hint: Think about whether notna() does anything isna() plus a negation couldn't already do.

No, there's no behavioral difference at all — notna() is implemented as exactly the logical negation of isna(), so notna() and manually negating isna()'s result always produce identical results for any input. notna() exists purely for readability, so code that's naturally about identifying present values doesn't need to express that as a double negative, not missing, when it can just say directly what it means.

Exercises

MediumPractice using df.notna() in a real scenario.

View Solution

import pandas as pd
import numpy as np

df = pd.DataFrame({"email": ["a@x.com", np.nan, "c@x.com"]})
print(df["email"].notna())

Frequently Asked Questions

Is there any actual behavioral difference between df.notna() and manually negating the result of df.isna()?▼

Related Functions

df-isnaboolean-indexingdf-dropna