How to Reorder DataFrame Column Order

In pandas, DataFrames consist of both rows and columns, with each column representing a separate feature or variable. The order of these columns is significant for data analysis and manipulation.

Problem: Rearranging Column Order

Consider the following DataFrame (df):

import numpy as np
import pandas as pd

df = pd.DataFrame(np.random.rand(10, 5))

After adding an additional column, let's say 'mean', via assignment:

df['mean'] = df.mean(1)

The objective is to move the 'mean' column to the front, making it the first column in the DataFrame, while preserving the order of the remaining columns.

Solution: Reordering Columns

An effective method to reorder columns is by re-assigning the DataFrame with a modified list of columns.

Step 1: Obtain Column List as Array

cols = df.columns.tolist()

This will result in an array containing the column names in their current order.

Step 2: Rearrange Column Order

Rearrange the column order within the list as desired. In this case, to move 'mean' to the beginning:

cols = cols[-1:]   cols[:-1]

Step 3: Reorder DataFrame Columns

Reorder the DataFrame using the rearranged column list:

df = df[cols]  # or df = df.ix[:, cols]