Randomly Select Columns from Pandas DataFrame

In this article, we will discuss how to randomly select columns from the Pandas Dataframe.

According to our requirement, we can randomly select columns from a pandas Database method where pandas df.sample() method helps us randomly select rows and columns.

Syntax of pandas sample() method:

Return a random selection of elements from an object’s axis. For repeatability, you may use the random_state parameter.

DataFrame.sample(n=None, frac=None, replace=False, weights=None, random_state=None, axis=None)

Parameters:

• n: int value, Number of random rows to generate.
• frac: Float value, Returns (float value * length of data frame values ). frac cannot be used with n.
• replace: Boolean value, return sample with replacement if True.
• random_state: int value or numpy.random.RandomState, optional. if set to a particular integer, will return same rows as sample in every iteration.
• axis: 0 or ‘row’ for Rows and 1 or ‘column’ for Columns.

Method 1: Select a single column at random

In this approach firstly the Pandas package is read with which the given CSV file is imported using pd.read_csv() method is used to read the dataset. df.sample() method is used to randomly select rows and columns. axis =’ columns’ says that we’re selecting columns. when “n” isn’t specified the method returns one random column by default.

To download the CSV file click here

Output:

Method 2: Select a number of columns at a random state

In this approach, If the user wants to select a certain number of columns more than 1 we use the parameter ‘n’ for this purpose. In the below example, we give n as 5. randomly selecting 5 columns from the database. 

Output:

Method 3: Allow a random selection of the same column more than once (by setting replace=True)

Here, in this approach, If the user wants to select a column more than once, or if repeatability is needed in our selection we should set the replace parameter to ‘True’ in the df.sample() method. Column ‘Bunkerfields’ is repeated twice.

Output:

Method 4: Select a portion of the total number of columns at random:

Here in this approach, if the user wants to select a portion of the dataset, the frac parameter should be used. In the below example our dataset has 10 columns. 0.25 of 10 is 2.5, it is further rounded to 2. A year and GasFlaring columns are returned. 

Output: