Pandas: How to Convert DataFrame to NumPy Array

Use the pandas DataFrame method .to_numpy() to convert a DataFrame to a NumPy array, like array = df.to_numpy().

Examples

Inputdf = pd.DataFrame({'A': [1, 2], 'B': [3, 4]})

Output[[1 3] [2 4]]

Inputdf = pd.DataFrame({'X': [5.5, 6.5], 'Y': [7.5, 8.5]})

Output[[5.5 7.5] [6.5 8.5]]

Inputdf = pd.DataFrame({'A': [1, None], 'B': [None, 4]})

Output[[ 1. nan] [nan 4.]]

How to Think About It

To convert a DataFrame to a NumPy array, think of extracting the raw data without the row and column labels. The .to_numpy() method returns the data as a plain array, which is useful for numerical operations or when working with libraries that require arrays.

Algorithm

1

Get the pandas DataFrame as input.

2

Call the <code>.to_numpy()</code> method on the DataFrame.

3

Return the resulting NumPy array.

Code

pandas

import pandas as pd

df = pd.DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})
array = df.to_numpy()
print(array)

Output

[[1 4]
 [2 5]
 [3 6]]

Dry Run

Let's trace converting a simple DataFrame to a NumPy array.

1

Create DataFrame

df = pd.DataFrame({'A': [1, 2], 'B': [3, 4]})

2

Call to_numpy()

array = df.to_numpy()

3

Resulting array

[[1 3] [2 4]]

A	B
1	3
2	4

Why This Works

Step 1: DataFrame stores data with labels

A pandas DataFrame holds data in rows and columns with labels for both.

Step 2: to_numpy() extracts raw data

The .to_numpy() method removes labels and returns just the data as a NumPy array.

Step 3: Result is easy for numerical work

The resulting array can be used in calculations or passed to libraries that need arrays.

Alternative Approaches

Using .values attribute

pandas

import pandas as pd

df = pd.DataFrame({'A': [1, 2], 'B': [3, 4]})
array = df.values
print(array)

This is an older way and works similarly but <code>.to_numpy()</code> is preferred for clarity and future compatibility.

Using numpy.array() on DataFrame

pandas

import pandas as pd
import numpy as np

df = pd.DataFrame({'A': [1, 2], 'B': [3, 4]})
array = np.array(df)
print(array)

This converts the DataFrame to a NumPy array but may be less explicit than <code>.to_numpy()</code>.

Complexity: O(n*m) time, O(n*m) space

Time Complexity

Converting a DataFrame to a NumPy array requires copying all data elements, so time grows with the number of rows (n) times columns (m).

Space Complexity

A new array is created to hold the data, so space also grows with n*m.

Which Approach is Fastest?

.to_numpy() and .values are similar in speed; using np.array() may add slight overhead.

Approach	Time	Space	Best For
.to_numpy()	O(n*m)	O(n*m)	Clear, recommended method
.values	O(n*m)	O(n*m)	Legacy code compatibility
np.array(df)	O(n*m)	O(n*m)	When working directly with NumPy