Split Pandas Dataframe by Column Index

Pandas support two data structures for storing data the series (single column) and dataframe where values are stored in a 2D table (rows and columns).  To index  a  dataframe using the index we need to make use of dataframe.iloc() method which takes 

Syntax: pandas.DataFrame.iloc[]

Parameters:
Index Position: Index position of rows in integer or list of integer.

Return type: Data frame or Series depending on parameters

Let's create a dataframe. In the below example we will use a simple binary dataset used to classify if a species is a mammal or reptile. The species column holds the labels where 1 stands for mammal and 0 for reptile. The data is stored in the dict which can be passed to the DataFrame function outputting a dataframe.

import pandas as pd

dataset = {'toothed': [1, 1, 1, 0, 1, 1, 1, 1, 1, 0],
           'hair': [1, 1, 0, 1, 1, 1, 0, 0, 1, 0],
           'breathes': [1, 1, 1, 1, 1, 1, 0, 1, 1, 1],
           'legs': [1, 1, 0, 1, 1, 1, 0, 0, 1, 1],
           'species': [1, 1, 0, 1, 1, 1, 0, 0, 1, 0]
           }

df = pd.DataFrame(dataset)

df.head()

Output :

Example 1: Now we would like to separate species columns from the feature columns (toothed, hair, breathes, legs) for this we are going to make use of the iloc[rows, columns] method offered by pandas. 

Here ':' stands for all the rows and -1 stands for the last column so the below cell is going to take the all the rows and all columns except the last one ('species') as can be seen in the output:

X = df.iloc[:,:-1]
X

Output: 

To split the species column from the rest of the dataset we make you of a similar code except in the cols position instead of padding a slice we pass in an integer value -1.

Y = df.iloc[:,-1]
Y

Output : 

Example 2: Splitting using list of integers 

Similar output can be obtained by passing in a list of integers instead of a slice 

X = df.iloc[:,[0,1,2,3]]
X

Output:

To the species column we are going to use the index of the column which is 4 we can use -1 as well 

Y = df.iloc[:,4]
Y

Output:

Example 3: Splitting dataframes into 2 separate dataframes 

In the above two examples, the output for Y was a Series and not a dataframe Now we are going to split the dataframe into two separate dataframe's this can be useful when dealing with multi-label datasets. Will be using the same dataset. 

In the first, we are going to split at column hair 

df.iloc[:,[0,1]] 

Output:

The second dataframe will contain 3 columns breathes , legs , species 

df.iloc[:,[2,3,4]] 

Output: