Why Left and right joins in Data Analysis Python? - Purpose & Use Cases

The Big Idea

What if you could combine messy lists perfectly without missing a single detail?

The Scenario

Imagine you have two lists of friends: one with their names and phone numbers, and another with their names and email addresses. You want to combine these lists to get a full contact list.

The Problem

Trying to match friends manually is slow and confusing. You might miss some friends who only appear in one list, or mix up their details. It's easy to make mistakes and waste time.

The Solution

Left and right joins let you combine these lists automatically. They keep all friends from one list and add matching info from the other, filling in missing details with blanks. This saves time and avoids errors.

Before vs After

✗ Before

combined = []
for friend in list1:
    for contact in list2:
        if friend['name'] == contact['name']:
            combined.append({**friend, **contact})

✓ After

combined = list1.merge(list2, how='left', on='name')

What It Enables

With left and right joins, you can easily combine data from different sources to get a complete picture without losing important information.

Real Life Example

A store wants to combine customer purchase records with their loyalty program info. Using a left join, they keep all purchases and add loyalty details when available.

Key Takeaways

Manual matching is slow and error-prone.

Left and right joins automate combining data based on shared keys.

They keep all data from one side and add matching info from the other.

Practice

1. What does a left join do when combining two tables in data analysis?

easy

A. Keeps all rows from both tables, matching or not.

B. Keeps all rows from the left table and adds matching rows from the right table.

C. Keeps only rows that match in both tables.

D. Keeps all rows from the right table and adds matching rows from the left table.

Solution

1. Step 1: Understand the role of the left table in a left join

   A left join keeps every row from the left table regardless of matches in the right table.

2. Step 2: Understand how matching rows from the right table are added

   Only rows from the right table that match the left table's keys are added; unmatched right rows are ignored.

3. Final Answer:

   Keeps all rows from the left table and adds matching rows from the right table. -> Option B

4. Quick Check:

   Left join = all left rows + matching right rows [OK]

Hint: Left join keeps all left rows, adds right matches [OK]

Common Mistakes:

• Confusing left join with right join
• Thinking left join keeps unmatched right rows
• Believing left join only keeps matching rows

2. Which of the following is the correct syntax for a left join using pandas in Python?

easy

A. pd.merge(df1, df2, how='left', on='key')

B. pd.merge(df1, df2, how='right', on='key')

C. pd.join(df1, df2, how='left')

D. df1.left_join(df2, on='key')

Solution

1. Step 1: Recall pandas merge syntax for joins

   In pandas, the function to join tables is pd.merge(), specifying the join type with the 'how' parameter.

2. Step 2: Identify the correct parameters for a left join

   Using how='left' and on='key' correctly performs a left join on the specified key column.

3. Final Answer:

   pd.merge(df1, df2, how='left', on='key') -> Option A

4. Quick Check:

   pd.merge + how='left' = left join [OK]

Hint: Use pd.merge with how='left' for left join [OK]

Common Mistakes:

• Using pd.join instead of pd.merge
• Wrong 'how' parameter like 'right' or missing
• Calling non-existent left_join method

3. Given these two DataFrames:

df1 = pd.DataFrame({'id': [1, 2, 3], 'value': ['A', 'B', 'C']})
df2 = pd.DataFrame({'id': [2, 3, 4], 'score': [90, 80, 70]})

What is the result of pd.merge(df1, df2, how='left', on='id')?

medium

A. [{'id':1,'value':'A','score':NaN},{'id':2,'value':'B','score':90},{'id':3,'value':'C','score':80}]

B. [{'id':2,'value':'B','score':90},{'id':3,'value':'C','score':80},{'id':4,'value':NaN,'score':70}]

C. [{'id':1,'value':'A','score':NaN},{'id':2,'value':'B','score':90},{'id':3,'value':'C','score':80},{'id':4,'value':NaN,'score':70}]

D. [{'id':1,'value':'A','score':NaN},{'id':2,'value':'B','score':NaN},{'id':3,'value':'C','score':NaN}]

Solution

1. Step 1: Identify the left and right tables

   df1 is the left table with ids 1, 2, 3; df2 is the right table with ids 2, 3, 4.

2. Step 2: Perform left join on 'id'

   All rows from df1 appear. For id=1, no match in df2, so score is NaN. For ids 2 and 3, scores 90 and 80 are added. id=4 from df2 is ignored.

3. Final Answer:

   [{'id':1,'value':'A','score':NaN},{'id':2,'value':'B','score':90},{'id':3,'value':'C','score':80}] -> Option A

4. Quick Check:

   Left join keeps all df1 rows + matching df2 rows [OK]

Hint: Left join keeps all left rows, unmatched right rows ignored [OK]

Common Mistakes:

• Including unmatched right rows in output
• Replacing unmatched left values with NaN incorrectly
• Confusing left and right join results

4. You run this code but get an error:

pd.merge(df1, df2, how='right', key='id')

What is the error and how to fix it?

medium

A. Error: DataFrames must have same columns. Fix by renaming columns.

B. Error: 'how' must be 'left' or 'inner'. Fix by changing 'right' to 'left'.

C. Error: 'key' is not a valid argument. Fix by using 'on='id'' instead.

D. No error, code runs fine.

Solution

1. Step 1: Identify the incorrect parameter name

   The parameter to specify join keys in pd.merge is 'on', not 'key'.

2. Step 2: Correct the parameter name

   Replace 'key='id'' with 'on='id'' to fix the error.

3. Final Answer:

   Error: 'key' is not a valid argument. Fix by using 'on='id'' instead. -> Option C

4. Quick Check:

   Use 'on' for join keys in pd.merge [OK]

Hint: Use 'on' not 'key' for join columns in pd.merge [OK]

Common Mistakes:

• Using 'key' instead of 'on' parameter
• Changing 'how' incorrectly
• Assuming no error occurs

5. You have two DataFrames:

df_customers = pd.DataFrame({'customer_id': [1, 2, 3], 'name': ['Ann', 'Bob', 'Cara']})
df_orders = pd.DataFrame({'order_id': [101, 102], 'customer_id': [2, 4], 'amount': [250, 300]})

You want to list all orders with customer names, including orders without matching customers. Which join should you use and why?

hard

A. Full outer join to keep all customers and all orders.

B. Left join on 'customer_id' to keep all customers and add matching orders.

C. Inner join on 'customer_id' to keep only matching customers and orders.

D. Right join on 'customer_id' to keep all orders and add matching customer names.

Solution

1. Step 1: Identify which table to keep all rows from

   You want all orders listed, even if no matching customer exists, so keep all rows from df_orders.

2. Step 2: Choose join type to keep all right table rows

   df_orders is the right table in the join. Using a right join keeps all rows from df_orders and adds matching customer names from df_customers.

3. Final Answer:

   Right join on 'customer_id' to keep all orders and add matching customer names. -> Option D

4. Quick Check:

   Right join keeps all right table rows [OK]

Hint: Use right join to keep all orders, add matching customers [OK]

Common Mistakes:

• Using left join keeps all customers, not all orders
• Using inner join drops unmatched orders
• Using full join adds unmatched customers unnecessarily