The desired data type of the returned NumPy array. By default, the data type will be the common type of the array's values. See examples below for clarification.

2. copylink | boolean | optional

If True, then a new NumPy array is created. Modifying this array would not affect the source DataFrame and vice versa.
If False, then a reference to the DataFrame's NumPy array representation is returned. This means that if you modify the array, then the original DataFrame will also be modified, and vice versa.

By default copy=False.

Return Value

A Numpy array holding all the values of the source DataFrame.

Examples

Obtaining the NumPy Array representation

Consider the following DataFrame:


        
        
            
                
                
                    df = pd.DataFrame({"A":[1,2], "B":[3,4]})
df
                
            
               A  B
0  1  3
1  2  4

To get the values of df as a NumPy array:


        
        
            
                
                
                    df.to_numpy()
                
            
            array([[1, 3],
       [2, 4]])

Data type of returned NumPy array

Consider the following DataFrame:


        
        
            
                
                
                    df = pd.DataFrame({"A":[1,2], "B":[3.0,4.0]})
df
                
            
               A  B
0  1  3.0
1  2  4.0

Here, column A is of type int, while column B is of type float.

The limitation with NumPy arrays is that all their values must be of one type. Since our df has two types, the to_numpy(~) method will opt to use float as int can be represented using float:


        
        
            
                
                
                    df.to_numpy().dtype
                
            
            dtype('float64')

Creating a new copy

Consider the following DataFrame:


        
        
            
                
                
                    df = pd.DataFrame({"A":[1,2], "B":[3,4]})
df
                
            
               A  B
0  1  3
1  2  4

To create a new NumPy array, set copy=True. In the code snippet below, we modify the first value of the array and check to see whether the source DataFrame, df, has been modified:


        
        
            
                
                
                    arr = df.to_numpy(copy=True)
arr[0,0] = 5
df
                
            
               A  B
0  1  3
1  2  4