Repeatable read behavior in PostgreSQL

Introduction

Repeatable read helps keep your data consistent during a transaction by making sure you see the same data each time you look, even if others change it.

When you want to run multiple queries in a transaction and need the data to stay the same throughout.

When you want to avoid seeing changes made by others while you work on your data.

When you want to prevent some types of data conflicts without locking the whole table.

When you want to ensure reports or calculations use stable data during their run.

Syntax

PostgreSQL

BEGIN TRANSACTION ISOLATION LEVEL REPEATABLE READ;
-- your queries here
COMMIT;

Use BEGIN TRANSACTION ISOLATION LEVEL REPEATABLE READ; to start a transaction with repeatable read isolation.

All queries inside this transaction will see the same snapshot of data.

Examples

This transaction reads the same data twice and will see no changes between the two SELECTs.

PostgreSQL

BEGIN TRANSACTION ISOLATION LEVEL REPEATABLE READ;
SELECT * FROM accounts WHERE user_id = 1;
SELECT * FROM accounts WHERE user_id = 1;
COMMIT;

Updates and reads happen in the same transaction, so the SELECT sees the updated balance.

PostgreSQL

BEGIN TRANSACTION ISOLATION LEVEL REPEATABLE READ;
UPDATE accounts SET balance = balance + 100 WHERE user_id = 1;
SELECT balance FROM accounts WHERE user_id = 1;
COMMIT;

Sample Program

This example shows that inside the repeatable read transaction, both SELECT queries see the same balance value, even if another session changes it between the queries.

PostgreSQL

CREATE TABLE accounts (user_id INT PRIMARY KEY, balance INT);
INSERT INTO accounts VALUES (1, 1000);

BEGIN TRANSACTION ISOLATION LEVEL REPEATABLE READ;
SELECT balance FROM accounts WHERE user_id = 1;
-- Imagine another session updates balance to 2000 here
SELECT balance FROM accounts WHERE user_id = 1;
COMMIT;

Important Notes

Repeatable read prevents non-repeatable reads but does not prevent phantom reads (new rows appearing).

It uses a snapshot of the database taken at the start of the transaction.

Other transactions can still change data, but you won't see those changes inside your transaction.

Summary

Repeatable read keeps your data consistent during a transaction by showing the same snapshot.

It helps avoid seeing changes made by others while you work.

Use it when you need stable data for multiple queries in one transaction.

Practice

1.

What does the REPEATABLE READ isolation level guarantee in PostgreSQL?

easy

A. It ensures all queries in a transaction see the same data snapshot.

B. It allows reading uncommitted changes from other transactions.

C. It locks all rows in the database for the transaction duration.

D. It automatically commits after each query in the transaction.

Solution

1. Step 1: Understand Repeatable Read isolation

   Repeatable Read ensures that all queries in a transaction see the same snapshot of data, preventing changes made by others from appearing during the transaction.

2. Step 2: Compare options with definition

   It ensures all queries in a transaction see the same data snapshot. matches this definition exactly. Options B, C, and D describe behaviors not related to Repeatable Read.

3. Final Answer:

   It ensures all queries in a transaction see the same data snapshot. -> Option A

4. Quick Check:

   Repeatable Read = Same snapshot [OK]

Hint: Repeatable Read = stable snapshot during transaction [OK]

Common Mistakes:

• Confusing Repeatable Read with Read Uncommitted
• Thinking it locks all rows
• Assuming auto-commit after each query

2.

Which of the following is the correct way to start a transaction with REPEATABLE READ isolation level in PostgreSQL?

BEGIN;
-- your queries
COMMIT;

easy

A. SET TRANSACTION ISOLATION LEVEL REPEATABLE READ; BEGIN;

B. BEGIN ISOLATION LEVEL REPEATABLE READ;

C. BEGIN; SET TRANSACTION ISOLATION LEVEL REPEATABLE READ;

D. BEGIN; SET SESSION CHARACTERISTICS AS TRANSACTION ISOLATION LEVEL REPEATABLE READ;

Solution

1. Step 1: Recall correct syntax for setting isolation level

   In PostgreSQL, you start the transaction with BEGIN, then set the isolation level for that transaction using SET TRANSACTION ISOLATION LEVEL.

2. Step 2: Match options to syntax

   BEGIN; SET TRANSACTION ISOLATION LEVEL REPEATABLE READ; correctly shows BEGIN; then SET TRANSACTION ISOLATION LEVEL REPEATABLE READ;. SET TRANSACTION ISOLATION LEVEL REPEATABLE READ; BEGIN; sets isolation before BEGIN which is invalid. BEGIN ISOLATION LEVEL REPEATABLE READ; uses invalid syntax. BEGIN; SET SESSION CHARACTERISTICS AS TRANSACTION ISOLATION LEVEL REPEATABLE READ; sets session level, not transaction level.

3. Final Answer:

   BEGIN; SET TRANSACTION ISOLATION LEVEL REPEATABLE READ; -> Option C

4. Quick Check:

   Set isolation after BEGIN = BEGIN; SET TRANSACTION ISOLATION LEVEL REPEATABLE READ; [OK]

Hint: Set isolation level after BEGIN with SET TRANSACTION [OK]

Common Mistakes:

• Trying to set isolation before BEGIN
• Using BEGIN with isolation level directly
• Confusing session and transaction level commands

3.

Consider this sequence in PostgreSQL with REPEATABLE READ isolation:

-- Transaction 1
BEGIN;
SELECT balance FROM accounts WHERE id = 1; -- returns 100

-- Transaction 2
BEGIN;
UPDATE accounts SET balance = 200 WHERE id = 1;
COMMIT;

-- Back to Transaction 1
SELECT balance FROM accounts WHERE id = 1;
COMMIT;

What will be the result of the second SELECT in Transaction 1?

medium

A. Error due to concurrent update

B. 100

C. NULL

D. 200

Solution

1. Step 1: Understand snapshot behavior in Repeatable Read

   Transaction 1 sees a consistent snapshot from its start. Changes committed by Transaction 2 after Transaction 1 began are not visible.

2. Step 2: Apply to the SELECT query

   The first SELECT returned 100. The second SELECT in the same transaction will also return 100, ignoring the update committed by Transaction 2.

3. Final Answer:

   100 -> Option B

4. Quick Check:

   Repeatable Read = same snapshot = 100 [OK]

Hint: Repeatable Read ignores later commits in same transaction [OK]

Common Mistakes:

• Expecting updated value 200 inside same transaction
• Thinking it causes error on concurrent update
• Assuming NULL if data changed

4.

Given this transaction in PostgreSQL:

BEGIN ISOLATION LEVEL REPEATABLE READ;
UPDATE products SET stock = stock - 1 WHERE id = 10;
SELECT stock FROM products WHERE id = 10;
COMMIT;

But you get an error: ERROR: syntax error at or near "ISOLATION". What is the fix?

medium

A. Change to BEGIN; SET TRANSACTION ISOLATION LEVEL REPEATABLE READ;

B. Remove the isolation level, just use BEGIN;

C. Use START TRANSACTION ISOLATION LEVEL REPEATABLE READ;

D. Use SET SESSION CHARACTERISTICS AS TRANSACTION ISOLATION LEVEL REPEATABLE READ; before BEGIN

Solution

1. Step 1: Identify syntax error cause

   PostgreSQL does not support specifying isolation level directly in BEGIN statement.

2. Step 2: Correct syntax to set isolation level

   You must first BEGIN; then run SET TRANSACTION ISOLATION LEVEL REPEATABLE READ; before queries.

3. Final Answer:

   Change to BEGIN; SET TRANSACTION ISOLATION LEVEL REPEATABLE READ; -> Option A

4. Quick Check:

   Set isolation after BEGIN, not inside [OK]

Hint: Set isolation after BEGIN with SET TRANSACTION [OK]

Common Mistakes:

• Trying to put isolation level inside BEGIN
• Using unsupported START TRANSACTION syntax
• Setting session level instead of transaction level

5.

You want to run multiple SELECT queries in a transaction and ensure the data does not change during the transaction, but you also want to allow other transactions to update data concurrently without blocking. Which isolation level should you choose in PostgreSQL?

hard

A. READ COMMITTED

B. READ UNCOMMITTED

C. SERIALIZABLE

D. REPEATABLE READ

Solution

1. Step 1: Understand isolation levels and concurrency

   READ COMMITTED allows seeing changes committed during the transaction, so data can change between queries. SERIALIZABLE is strict and may block or abort concurrent updates. READ UNCOMMITTED is not supported in PostgreSQL.

2. Step 2: Match requirement to isolation level

   REPEATABLE READ provides a stable snapshot for all queries in the transaction, preventing data changes from appearing, while allowing concurrent updates without blocking reads.

3. Final Answer:

   REPEATABLE READ -> Option D

4. Quick Check:

   Stable snapshot + concurrency = REPEATABLE READ [OK]

Hint: Repeatable Read = stable snapshot without blocking writes [OK]

Common Mistakes:

• Choosing SERIALIZABLE which blocks more
• Thinking READ COMMITTED prevents data changes
• Assuming READ UNCOMMITTED exists in PostgreSQL