Exception handling (BEGIN-EXCEPTION-END) in PostgreSQL - Mini Project: Build & Apply

Handling Errors in PostgreSQL with BEGIN-EXCEPTION-END

📖 Scenario: You are managing a small bookstore database. Sometimes, when adding new books, errors can happen, like trying to add a book with a duplicate ID. You want to handle these errors gracefully so the database stays safe and you get a clear message about what went wrong.

🎯 Goal: Build a simple PostgreSQL function that tries to insert a new book into the books table. Use BEGIN-EXCEPTION-END blocks to catch errors like duplicate IDs and handle them by returning a friendly message.

📋 What You'll Learn

Create a books table with columns id (integer primary key) and title (text).

Write a function called add_book that takes book_id and book_title as inputs.

Use a BEGIN-EXCEPTION-END block inside the function to try inserting the book.

Catch the unique_violation exception when a duplicate id is inserted.

Return a message indicating success or the error caught.

💡 Why This Matters

🌍 Real World

Handling errors in database operations is important to keep data safe and provide clear feedback to users or applications.

💼 Career

Database developers and administrators often write functions with exception handling to make systems robust and user-friendly.

1

Create the books table

Create a table called books with two columns: id as an integer primary key and title as text.

PostgreSQL

-- Your code here to create the books table

Hint

Use CREATE TABLE with id INTEGER PRIMARY KEY and title TEXT.

2

Start the add_book function

Write the first line of a function called add_book that takes two parameters: book_id as integer and book_title as text. The function should return text.

PostgreSQL

CREATE TABLE books (
    id INTEGER PRIMARY KEY,
    title TEXT
);

-- Start the add_book function definition here

Hint

Use CREATE OR REPLACE FUNCTION with parameters and RETURNS TEXT. Start the BEGIN block.

3

Add the insert and exception block

Inside the BEGIN block of add_book, write an INSERT statement to add the book to books. Then add an EXCEPTION block to catch unique_violation errors and return the message 'Duplicate book ID'.

PostgreSQL

CREATE TABLE books (
    id INTEGER PRIMARY KEY,
    title TEXT
);

CREATE OR REPLACE FUNCTION add_book(book_id INTEGER, book_title TEXT) RETURNS TEXT AS $$
BEGIN
    -- Insert the book here
    -- Add EXCEPTION block to catch unique_violation
END;
$$ LANGUAGE plpgsql;

Hint

Use INSERT INTO inside BEGIN. Add EXCEPTION WHEN unique_violation THEN to catch duplicates.

4

Complete the function with success message

Make sure the function add_book returns 'Book added successfully' when the insert works without errors.

PostgreSQL

CREATE TABLE books (
    id INTEGER PRIMARY KEY,
    title TEXT
);

CREATE OR REPLACE FUNCTION add_book(book_id INTEGER, book_title TEXT) RETURNS TEXT AS $$
BEGIN
    INSERT INTO books(id, title) VALUES (book_id, book_title);
    -- Ensure success message is returned
EXCEPTION
    WHEN unique_violation THEN
        RETURN 'Duplicate book ID';
END;
$$ LANGUAGE plpgsql;

Hint

Return the success message right after the insert statement inside BEGIN.

Practice

1. What is the main purpose of using a BEGIN-EXCEPTION-END block in PostgreSQL?

easy

A. To catch and handle errors during code execution

B. To speed up query execution

C. To create new tables automatically

D. To backup the database

Solution

1. Step 1: Understand the role of BEGIN-EXCEPTION-END

   This block is designed to catch errors that happen during the execution of SQL or PL/pgSQL code.

2. Step 2: Identify the purpose of error handling

   It allows the code to handle errors gracefully without stopping the entire process.

3. Final Answer:

   To catch and handle errors during code execution -> Option A

4. Quick Check:

   Error handling = A [OK]

Hint: Remember: BEGIN-EXCEPTION-END is for error catching [OK]

Common Mistakes:

• Thinking it speeds up queries
• Confusing it with table creation
• Assuming it backs up data

2. Which of the following is the correct syntax to start an exception block in PostgreSQL?

easy

A. BEGIN TRY ... EXCEPTION ... END

B. START ... EXCEPTION ... FINISH

C. BEGIN ... EXCEPTION ... END

D. TRY ... CATCH ... END

Solution

1. Step 1: Recall PostgreSQL exception block syntax

   PostgreSQL uses BEGIN ... EXCEPTION ... END to define a block with error handling.

2. Step 2: Compare options with correct syntax

   Only BEGIN ... EXCEPTION ... END matches the exact syntax used in PostgreSQL.

3. Final Answer:

   BEGIN ... EXCEPTION ... END -> Option C

4. Quick Check:

   PostgreSQL uses BEGIN-EXCEPTION-END syntax [OK]

Hint: Look for BEGIN and EXCEPTION keywords together [OK]

Common Mistakes:

• Using TRY-CATCH like other languages
• Adding extra keywords like START or FINISH
• Mixing syntax from other SQL dialects

3. What will be the output of the following PostgreSQL block?

DO $$
BEGIN
  RAISE EXCEPTION 'Error happened';
EXCEPTION
  WHEN OTHERS THEN
    RAISE NOTICE 'Caught an error';
END $$;

medium

A. The block raises an error and stops execution

B. It prints 'Caught an error' as a notice and continues

C. It prints 'Error happened' as a notice

D. Syntax error due to missing semicolon

Solution

1. Step 1: Analyze the RAISE EXCEPTION statement

   The block raises an exception with message 'Error happened'.

2. Step 2: Check the EXCEPTION block handling

   The exception handler catches all errors (WHEN OTHERS) and raises a NOTICE with 'Caught an error'.

3. Final Answer:

   It prints 'Caught an error' as a notice and continues -> Option B

4. Quick Check:

   Exception caught triggers NOTICE message [OK]

Hint: WHEN OTHERS catches all errors and can raise NOTICE [OK]

Common Mistakes:

• Assuming the error stops execution
• Confusing EXCEPTION message with NOTICE
• Thinking syntax error occurs

4. Identify the error in this PostgreSQL exception block:

BEGIN
  PERFORM 1/0;
EXCEPTION
  WHEN division_by_zero THEN
    RAISE NOTICE 'Division by zero caught';
END;

medium

A. Missing DO $$ ... $$ wrapper for anonymous block

B. Incorrect exception name; should be zero_division_error

C. RAISE NOTICE cannot be used inside EXCEPTION

D. No error; block is correct

Solution

1. Step 1: Check block context

   In PostgreSQL, standalone BEGIN-EXCEPTION-END blocks must be inside a DO block or function.

2. Step 2: Verify exception name and usage

   Exception name 'division_by_zero' is valid, and RAISE NOTICE is allowed inside EXCEPTION.

3. Step 3: Identify missing DO wrapper

   The code lacks DO $$ ... $$ to run as an anonymous block.

4. Final Answer:

   Missing DO $$ ... $$ wrapper for anonymous block -> Option A

5. Quick Check:

   Anonymous blocks need DO wrapper [OK]

Hint: Use DO $$ ... $$ for anonymous BEGIN blocks [OK]

Common Mistakes:

• Thinking exception name is wrong
• Believing RAISE NOTICE is invalid here
• Ignoring the need for DO wrapper

5. You want to write a PostgreSQL block that tries to insert a row into a table but if a unique constraint violation occurs, it should print 'Duplicate found' and continue without error. Which code snippet correctly implements this?

hard

A. BEGIN
INSERT INTO users(id) VALUES(1);
EXCEPTION
WHEN duplicate_key THEN
RAISE NOTICE 'Duplicate found';
END;

B. BEGIN
INSERT INTO users(id) VALUES(1);
EXCEPTION
WHEN unique_error THEN
RAISE NOTICE 'Duplicate found';
END;

C. BEGIN
INSERT INTO users(id) VALUES(1);
EXCEPTION
WHEN unique_constraint THEN
RAISE NOTICE 'Duplicate found';
END;

D. BEGIN
INSERT INTO users(id) VALUES(1);
EXCEPTION
WHEN unique_violation THEN
RAISE NOTICE 'Duplicate found';
END;

Solution

1. Step 1: Identify correct exception name for unique constraint

   PostgreSQL uses unique_violation to catch unique constraint errors.

2. Step 2: Check the exception block syntax

   The block catches unique_violation and raises a notice 'Duplicate found' correctly.

3. Step 3: Verify other options

   Other options use incorrect exception names that do not exist in PostgreSQL.

4. Final Answer:

   BEGIN INSERT INTO users(id) VALUES(1); EXCEPTION WHEN unique_violation THEN RAISE NOTICE 'Duplicate found'; END; -> Option D

5. Quick Check:

   Use unique_violation for unique constraint errors [OK]

Hint: Use unique_violation to catch duplicate key errors [OK]

Common Mistakes:

• Using wrong exception names
• Forgetting to handle unique_violation
• Not raising a notice or message