🏗️ Reshaping Your Data Table: A Builder’s Guide

Imagine your DataFrame is a LEGO table. You can add new columns of blocks, remove pieces you don’t need, and rearrange everything exactly how you want. Let’s become master builders!

🧱 The Big Picture

Think of a DataFrame like a spreadsheet or a toy organizer with slots. Sometimes you need to:

• Add a new slot (column) for a new type of toy
• Remove a slot you don’t use anymore
• Rearrange where things go
• Make a copy so you can experiment without breaking the original

Let’s learn each skill, one by one!

➕ Adding a New Column

The Story: Your toy box has columns for “Toy Name” and “Color”. But now you want to track “Price” too!

It’s Super Easy:

# Your toy box
df = pd.DataFrame({
    'Toy': ['Car', 'Ball', 'Doll'],
    'Color': ['Red', 'Blue', 'Pink']
})

# Add a new column - just like this!
df['Price'] = [10, 5, 15]

What Happened?

✨ Magic Trick: You can also add a column where every row has the same value:

df['InStock'] = True  # All toys are in stock!

🗑️ Deleting Columns with drop()

The Story: You decided you don’t want to track “Color” anymore. Time to remove that column!

# Remove ONE column
df = df.drop('Color', axis=1)

# Remove MULTIPLE columns at once
df = df.drop(['Color', 'Price'], axis=1)

The Secret Code: axis=1 means “columns” (think: columns go sideways, like the number 1 lying down)

graph TD
    A[Original Table] --> B{drop with axis=1}
    B --> C[Column Removed!]
    style C fill:#90EE90

⚠️ Important: By default, drop() gives you a NEW table. The original stays the same. Want to change the original? Add inplace=True:

df.drop('Color', axis=1, inplace=True)

🚮 Deleting Rows with drop()

The Story: That broken toy in row 1? Let’s remove it!

# Remove row at index 1
df = df.drop(1, axis=0)

# Remove multiple rows
df = df.drop([0, 2], axis=0)

The Secret Code: axis=0 means “rows” (think: rows go up-down, like you standing up = 0)

🎪 The pop() Method - Remove AND Keep

The Story: What if you want to remove a column BUT also save it somewhere else? Like taking a toy out of the box to give to a friend!

# Pop removes the column AND returns it
saved_prices = df.pop('Price')

print(saved_prices)
# Output: [10, 5, 15]

Pop vs Drop:

⚠️ Note: pop() only works for columns, not rows!

📍 Inserting Columns at a Specific Spot

The Story: You want to add “Weight” as the SECOND column, not at the end!

# insert(position, name, values)
df.insert(1, 'Weight', [2, 1, 3])

Before: | Toy | Color | Price |

After inserting at position 1: | Toy | Weight | Color | Price |

✨ Position 0 = first column, Position 1 = second column, and so on!

🎁 The assign() Method - Add Without Changing

The Story: Imagine you want to TRY adding a column, but you don’t want to mess up your original table. Like trying on clothes without buying them!

# Creates a NEW table with the extra column
new_df = df.assign(Discount=0.10)

# Original df is UNCHANGED!
# new_df has the 'Discount' column

Super Power: Add multiple columns at once!

new_df = df.assign(
    Discount=0.10,
    FinalPrice=df['Price'] * 0.90
)

graph TD
    A[Original DataFrame] --> B[assign method]
    B --> C[NEW DataFrame with extra columns]
    A --> D[Original stays SAFE]
    style C fill:#87CEEB
    style D fill:#90EE90

💥 The explode() Method - Unpacking Lists

The Story: What if one cell has MULTIPLE values inside it? Like a toy box that contains a bag with 3 marbles?

df = pd.DataFrame({
    'Kid': ['Alice', 'Bob'],
    'Toys': [['Car', 'Ball'], ['Doll']]
})

Use explode() to unpack:

df = df.explode('Toys')

🎆 Each item in the list gets its own row!

📋 Copying DataFrames - The Safe Way

The Story: You want to experiment with your table, but you’re scared of breaking it. Make a copy first!

⚠️ The WRONG Way (Shallow Copy)

copy_df = df  # This is NOT a real copy!

If you change copy_df, you change df too! They’re the same table with two names.

✅ The RIGHT Way (Deep Copy)

copy_df = df.copy()  # This IS a real copy!

Now copy_df is completely separate. Change it all you want!

graph TD
    A[Original df] --> B{How to copy?}
    B -->|df2 = df| C[SAME table, 2 names]
    B -->|df2 = df.copy| D[TWO separate tables]
    style C fill:#FFB6C1
    style D fill:#90EE90

🎯 Quick Reference Card

🌟 You Did It!

You now know how to:

• ➕ Add new columns anywhere you want
• 🗑️ Remove columns or rows you don’t need
• 🎪 Pop columns to use elsewhere
• 📍 Insert columns at exact positions
• 🎁 Safely add columns without changing the original
• 💥 Explode lists into separate rows
• 📋 Make safe copies of your data

You’re a DataFrame architect now! 🏆