Pandas DataFrame is a two-dimensional size-mutable, potentially heterogeneous tabular data structure with labeled axes (rows and columns). This data structure can be converted to NumPy ndarray with the help of the DataFrame.to_numpy() method. In this article we will see how to convert dataframe to numpy array.
Syntax of Pandas DataFrame.to_numpy()
Syntax: Dataframe.to_numpy(dtype = None, copy = False)
Parameters:
- dtype: Data type which we are passing like str.
- copy: [bool, default False] Ensures that the returned value is a not a view on another array.
Returns: numpy.ndarray
Convert DataFrame to Numpy Array
Here, we will see how to convert DataFrame to a Numpy array.
Python3
import pandas as pd# initialize a dataframedf = pd.DataFrame( [[1, 2, 3], [4, 5, 6], [7, 8, 9], [10, 11, 12]], columns=['a', 'b', 'c'])# convert dataframe to numpy arrayarr = df.to_numpy()print('\nNumpy Array\n----------\n', arr)print(type(arr)) |
Output:
Numpy Array ---------- [[ 1 2 3] [ 4 5 6] [ 7 8 9] [10 11 12]] <class 'numpy.ndarray'>
Here we want to convert a particular column into numpy array.
Python3
import pandas as pd# initialize a dataframedf = pd.DataFrame( [[1, 2, 3], [4, 5, 6], [7, 8, 9], [10, 11, 12]], columns=['a', 'b', 'c'])# convert dataframe to numpy arrayarr = df[['a', 'c']].to_numpy()print('\nNumpy Array\n----------\n', arr)print(type(arr)) |
Output:
Numpy Array ---------- [[ 1 3] [ 4 6] [ 7 9] [10 12]] <class 'numpy.ndarray'>
Here we are converting a dataframe with different datatypes.
Python3
import pandas as pdimport numpy as np#initialize a dataframedf = pd.DataFrame( [[1, 2, 3], [4, 5, 6.5], [7, 8.5, 9], [10, 11, 12]], columns=['a', 'b', 'c'])arr = df.to_numpy()print('Numpy Array', arr)print('Numpy Array Datatype :', arr.dtype) |
Output:
Numpy Array [[ 1. 2. 3. ] [ 4. 5. 6.5] [ 7. 8.5 9. ] [10. 11. 12. ]] Numpy Array Datatype : float64
To get the link to the CSV file, click on nba.csv
Example 1:
Here, we are using a CSV file for changing the Dataframe into a Numpy array by using the method DataFrame.to_numpy(). After that, we are printing the first five values of the Weight column by using the df.head() method.
Python3
# importing pandasimport pandas as pd# reading the csvdata = pd.read_csv("nba.csv")data.dropna(inplace=True)# creating DataFrame from weight columndf = pd.DataFrame(data['Weight'].head())# using to_numpy() functionprint(df.to_numpy()) |
Output:
[[180.] [235.] [185.] [235.] [238.]]
Example 2:
In this example, we are just providing the parameters in the same code to provide the dtype here.
Python3
# importing pandasimport pandas as pd# read csv filedata = pd.read_csv("nba.csv")data.dropna(inplace=True)# creating DataFrame from weight columndf = pd.DataFrame(data['Weight'].head())# providing dtypeprint(df.to_numpy(dtype='float32')) |
Output:
[[180.] [235.] [185.] [235.] [238.]]
Example 3:
Validating the type of the array after conversion.
Python3
# importing pandasimport pandas as pd# reading csvdata = pd.read_csv("nba.csv")data.dropna(inplace=True)# creating DataFrame from weight columndf = pd.DataFrame(data['Weight'].head())# using to_numpy()print(type(df.to_numpy())) |
Output:
<class 'numpy.ndarray'>
