Pandas Series and DataFrames

# CHAPTER 14

Pandas Series and DataFrames #

1. Chapter Introduction #

2. Selecting Columns #

When you select a single column from a DataFrame, Pandas returns a 1D Series. If you select multiple columns, it returns a smaller 2D DataFrame.

import pandas as pd

data = {
    "Name": ["Alice", "Bob", "Charlie", "Dave"],
    "Age": [25, 30, 35, 40],
    "City": ["NY", "LA", "NY", "SF"]
}
df = pd.DataFrame(data)

# 1. Select a SINGLE column (Returns a Series)
# Syntax: df['ColumnName']
ages = df['Age']
print(type(ages)) # <class 'pandas.core.series.Series'>

# 2. Select MULTIPLE columns (Returns a DataFrame)
# Syntax: df[['Col1', 'Col2']]  <-- Notice the double brackets!
subset = df[[&#039;Name', 'City']]
print(subset.head(2))

3. Filtering Rows (Boolean Indexing) #

Filtering in Pandas works exactly like Boolean Indexing in NumPy. You pass a condition inside the brackets.

# 1. Find everyone older than 28
over_28 = df[df['Age'] > 28]
print(over_28)

# 2. Find everyone who lives in NY
ny_residents = df[df['City'] == 'NY']

# 3. Complex Filtering (AND = &, OR = |)
# Parentheses are REQUIRED around each condition!
target = df[(df['Age'] > 28) & (df['City'] == 'NY')]
print("Older than 28 AND in NY:\n", target)

4. Row Selection using .loc and .iloc #

How do you get the 3rd row of the table? Pandas provides two special indexers.

• .iloc (Integer Location): Selects by absolute numerical position (0, 1, 2...).

• .loc (Label Location): Selects by the index *name* (useful if your rows are named 'RowA', 'RowB' instead of 0, 1).

# Select the 1st row (Index 0) using integer location
first_row = df.iloc[0]

# Select rows 0 to 2, and columns 0 to 1
slice_df = df.iloc[0:2, 0:2]

5. Creating and Modifying Columns #

Creating a new column is as easy as declaring a new dictionary key. Because Pandas is built on NumPy, calculations are vectorized (no loops needed!).

# Create a new column by applying math to an existing one
df['Age_In_10_Years'] = df['Age'] + 10

# Create a boolean flag column
df['Is_Senior'] = df['Age'] >= 40

print(df)

6. Sorting Data #

You can quickly sort the DataFrame by any column.

# Sort by Age descending (oldest first)
sorted_df = df.sort_values(by='Age', ascending=False)
print(sorted_df)

7. Mini Project: Sales Filter #

# Simulating sales data
sales_data = {
    "OrderID": [1, 2, 3, 4],
    "Product": ["Laptop", "Mouse", "Monitor", "Keyboard"],
    "Price": [1200, 25, 300, 50],
    "Quantity": [2, 5, 1, 3]
}
sales = pd.DataFrame(sales_data)

# 1. Create a Total_Revenue column
sales['Total_Revenue'] = sales['Price'] * sales['Quantity']

# 2. Filter for orders bringing in more than $100
big_orders = sales[sales['Total_Revenue'] > 100]

# 3. Sort by biggest revenue
final_report = big_orders.sort_values(by='Total_Revenue', ascending=False)
print(final_report)

8. Common Mistakes #

• Forgetting Double Brackets: df['Name', 'Age'] will crash. To pass a list of multiple columns, you must use double brackets: df[['Name', 'Age']].

• Using and instead of &: In standard Python, you use the word and. In Pandas filters, you MUST use the bitwise operator & (and | for or), and you MUST wrap conditions in parentheses: (df['A'] > 1) & (df['B'] < 5).

9. MCQs #

10. Interview Questions #

• Q: Explain the difference between .loc and .iloc.

• Q: I have a DataFrame df. Why does df['Age'] > 30 return a Series of True/False values, while df[df['Age'] > 30] returns a DataFrame?

11. Summary #

12. Next Chapter Recommendation #