How to Add and Remove Rows and Columns in Pandas DataFrames

Introduction

In this article we walk through the most common pandas operations for modifying the shape of a DataFrame: adding and dropping columns, combining column values, and inserting or removing rows. The examples use a small "people" DataFrame with first name, last name, and email columns.

Adding Columns

• Create a new Series by concatenating existing columns: python full_name = df["first_name"] + " " + df["last_name"]
• Assign the Series to a new column using bracket notation (dot notation does not work for assignment): python df["full_name"] = full_name
• You can also generate a column with apply for more complex calculations.

Removing Columns

• Use drop with columns= and a list of column names: python df.drop(columns=["first_name", "last_name"], inplace=True)
• inplace=True makes the change permanent; otherwise drop returns a new DataFrame.

Splitting a Column Back into Two

• Apply the string split method with expand=True to turn a single column into two separate columns: python df[["first_name", "last_name"]] = df["full_name"].str.split(" ", expand=True)
• The left‑hand side uses a list of new column names inside brackets.

Adding Rows

Single Row

• append a dictionary (or Series) and set ignore_index=True so pandas creates a new index automatically: python df = df.append({"first_name": "Tony"}, ignore_index=True)
• Missing columns are filled with NaN.

Appending Another DataFrame

• Create a second DataFrame (df2) and append it: python df = df.append(df2, ignore_index=True, sort=False)
• sort=False suppresses the warning about column ordering.
• Like column drops, append does not have an inplace argument; assign the result back to df.

Removing Rows

• Drop by index: python df.drop(index=4, inplace=True)
• Conditional drop using a boolean filter: python filt = df["last_name"] == "Doe" df.drop(index=df[filt].index, inplace=True)
• Storing the filter in a variable improves readability.

Quick Recap

• Add column → create Series → df["new"] = series
• Remove column → df.drop(columns=[...], inplace=True)
• Split column → df[["col1","col2"]] = df["col"].str.split(" ", expand=True)
• Add row → df.append(dict, ignore_index=True) (or pd.concat)
• Remove row → df.drop(index=..., inplace=True) or conditional filter.

Why These Operations Matter

Manipulating the structure of a DataFrame is essential for cleaning raw data, preparing features for analysis, and merging datasets. Mastery of these basics speeds up any data‑science workflow.

By mastering df["col"] = …, drop, str.split, and append (or concat), you can reshape pandas DataFrames on the fly, making data cleaning and feature engineering far more efficient.