How To Change Column Order Using Pandas

Filed Under: Pandas
Change Column Order Using Pandas

In this article, we’ll learn how to change column order using Pandas. Pandas is an invaluable part of the data science world. For all your data manipulation and analysis, Pandas offers many amazing functions which can help you in the process. Pandas support dataframe objects to store data that has labeled rows and columns.

As you already know, data.columns function can list out all the columns / variable names in your data. But, how can you change the order of the columns?. Well, it’s an interesting question and I got multiple methods for it. 

So, without spending much time on Pandas, let’s see 4 different methods, using which you can change column order in python.


Change Column Order Using Pandas

As a first step, we have to import the required libraries for this purpose. We need Numpy and Pandas to work with data and our data will be a “titanic” dataset. 

#import libraries

import numpy as np
import pandas as pd

Let’s load the data using the Pandas read_csv() function.

#data

import pandas as pd

data = pd.read_csv('titanic.csv')
Titanic

Here is our Titanic dataset. Now, we are going to print out the column / variable names in this data as a list.

#columns

data.columns
Index(['PassengerId', 'Survived', 'Pclass', 'Name', 'Sex', 'Age', 'SibSp',
       'Parch', 'Ticket', 'Fare', 'Cabin', 'Embarked'],
      dtype='object')

Also check for duplicate columns / variables.

#check for duplicates 

data.columns.value_counts()
PassengerId    1
Fare           1
Embarked       1
Sex            1
Ticket         1
Pclass         1
Age            1
Survived       1
Parch          1
Name           1
Cabin          1
SibSp          1
dtype: int64

So, we don’t have any duplicate columns in our data. We are good to go 🙂


1. Pandas iloc Method

Using the Pandas iloc method, you can index or change column order in a specified order as shown below.

#iloc method

data.iloc[:, [3,5,4,9,2]]
Column order
  • It may look like a VLOOKUP table but it’s not. As shown above, you can specify the order of the column to arrange them as shown.You can play with a different order as per your use case.

Also read: Pandas Indexing: loc, iloc, and ix in Python


2. Pandas loc Method

Yes, using pandas loc method also, you can change column order in the data. Let’s see how it works!

#loc method

data.loc[:, ['Name','Age','Sex','Fare','Pclass']]
Column order

This pandas loc method also produces the same output. But, take some time and observe the difference between the working nature of these 2 methods.

In the iloc method, we specify integer input. But, in the loc method, you can pass both label and integer input. I have added an informative picture of the difference between the iloc and loc methods.

iloc vs loc
Source – Image by B. Chen
  • Above, I have mentioned the same order as the iloc method to show the working difference of them as well. Feel free to change the order and get your hands dirty with your data.

3. Pandas Subset Method

The pandas’ sub-setting method is one of the simplest methods among the above methods. You have to subset the data with a required order.

#subset

data[['Name','Age','Sex','Fare','Pclass']]
Column order

Well, we got the desired output. Just like this, you can subset the data with a use-case-specific order to get a newly ordered dataframe as shown above.

Note that you can not only order the columns but also slice them and extract the required data.


4. Pandas Reverse

The final method is using the pandas reverse. But, I don’t think it can be a very useful method. Because this method will just reverse the order of the data 😛

Example;

A -> Z,

Z -> A

#revese

rev_columns = list(data.columns)
rev_columns.reverse()
data[cols]
Column order

Well, we have reversed the order of all the columns. This is how it works! Let me know your thoughts on this method in the comments!


Change Column Order In Pandas – Wrapping Up

Sometimes you may need to change the column order of your data for a use case and you can use any of the above-shown methods based on your requirement. Pandas offer many functions, which help amazingly in our data analysis and wrangling. I have covered 4 methods in this story and probably they can be handy at some time.

That’s all for now! Happy Python!!! 🙂

More read: Pandas documentation (iloc)

Leave a Reply

Your email address will not be published. Required fields are marked *

close
Generic selectors
Exact matches only
Search in title
Search in content