Why Multi-key transactions for consistency in Redis? - Purpose & Use Cases

The Big Idea

What if your data could never get mixed up, no matter how many changes happen at once?

The Scenario

Imagine you are managing a shared piggy bank with your friends. Each friend can add or take money, but you write down every change on a piece of paper. If two friends write at the same time, the notes get mixed up and the total money count becomes wrong.

The Problem

Writing changes manually is slow and confusing. Mistakes happen easily when multiple people update the piggy bank at once. You might lose track of who added or took money, causing errors and disagreements.

The Solution

Multi-key transactions let you group all changes together and apply them at once. This way, either all updates happen or none do, keeping the piggy bank's total correct and consistent no matter how many friends update it simultaneously.

Before vs After

✗ Before

SET balance1 100
SET balance2 200
-- separate commands, risk of partial updates

✓ After

MULTI
INCRBY balance1 50
DECRBY balance2 50
EXEC
-- all changes applied together atomically

What It Enables

It enables safe, reliable updates across multiple keys, ensuring your data stays accurate and trustworthy even with many changes happening at once.

Real Life Example

In an online game, when a player trades items with another, multi-key transactions ensure both players' inventories update together perfectly, so no items disappear or duplicate.

Key Takeaways

Manual updates can cause errors when done separately.

Multi-key transactions group changes to keep data consistent.

This ensures reliable and accurate updates in shared data.

Practice

1. What is the main purpose of using MULTI and EXEC commands in Redis transactions?

easy

A. To group multiple commands so they execute atomically for data consistency

B. To speed up the execution of individual Redis commands

C. To create backups of the Redis database automatically

D. To delete multiple keys at once

Solution

1. Step 1: Understand MULTI command

   The MULTI command starts a transaction block in Redis, grouping commands.

2. Step 2: Understand EXEC command

   The EXEC command executes all commands queued after MULTI atomically, ensuring consistency.

3. Final Answer:

   To group multiple commands so they execute atomically for data consistency -> Option A

4. Quick Check:

   Atomic execution = Data consistency [OK]

Hint: MULTI starts, EXEC runs all commands atomically [OK]

Common Mistakes:

• Thinking MULTI speeds up commands individually
• Confusing transactions with backups
• Assuming EXEC deletes keys

2. Which of the following is the correct syntax to start a Redis transaction?

easy

A. EXEC

B. START

C. BEGIN

D. MULTI

Solution

1. Step 1: Identify Redis transaction start command

   Redis uses MULTI to begin a transaction block.

2. Step 2: Check other options

   EXEC ends the transaction, BEGIN and START are not Redis commands.

3. Final Answer:

   MULTI -> Option D

4. Quick Check:

   Start transaction = MULTI [OK]

Hint: Start with MULTI, end with EXEC [OK]

Common Mistakes:

• Using EXEC to start instead of end
• Using SQL-like BEGIN or START commands
• Confusing transaction commands with key commands

3. Consider the following Redis commands executed in a transaction:

MULTI
SET key1 "value1"
SET key2 "value2"
EXEC

What will be the state of key1 and key2 after EXEC?

medium

A. key1 and key2 will both have their new values set atomically

B. Only key1 will be set, key2 will remain unchanged

C. Neither key1 nor key2 will be set because EXEC fails

D. key1 will be set, but key2 will be deleted

Solution

1. Step 1: Understand MULTI and EXEC behavior

   Commands between MULTI and EXEC are queued and executed atomically.

2. Step 2: Analyze the commands

   Both SET key1 "value1" and SET key2 "value2" will be applied together when EXEC runs.

3. Final Answer:

   key1 and key2 will both have their new values set atomically -> Option A

4. Quick Check:

   Atomic set of keys = Both keys updated [OK]

Hint: EXEC applies all queued commands together [OK]

Common Mistakes:

• Thinking commands run immediately after MULTI
• Assuming partial execution on EXEC failure
• Confusing EXEC with DISCARD

4. You wrote this Redis transaction:

MULTI
SET key1 "val1"
GET key1
EXEC

But the GET key1 command returns null inside the transaction. Why?

medium

A. SET command failed, so GET returns null

B. GET command syntax is incorrect

C. GET inside MULTI queues the command and returns null until EXEC runs

D. EXEC was not called, so commands did not run

Solution

1. Step 1: Understand command queuing in MULTI

   Inside a transaction, commands are queued and do not execute immediately.

2. Step 2: Effect on GET command

   GET returns null because the SET has not executed yet; actual results come after EXEC.

3. Final Answer:

   GET inside MULTI queues the command and returns null until EXEC runs -> Option C

4. Quick Check:

   Queued commands return null before EXEC [OK]

Hint: GET returns null inside MULTI before EXEC [OK]

Common Mistakes:

• Thinking GET runs immediately inside MULTI
• Assuming SET failed without checking
• Forgetting to call EXEC

5. You want to update two related keys account:balance and account:points atomically in Redis. Which approach ensures both keys update together or not at all?

hard

A. Use GET to check keys, then SET commands separately

B. Use MULTI, queue both SET commands, then EXEC to apply atomically

C. Use WATCH on both keys, then run SET commands without MULTI

D. Run two separate SET commands without MULTI

Solution

1. Step 1: Understand atomic updates need transactions

   To update multiple keys together, Redis transactions with MULTI and EXEC ensure atomicity.

2. Step 2: Analyze other options

   Separate SET commands risk partial updates; WATCH helps optimistic locking but needs MULTI for atomicity.

3. Final Answer:

   Use MULTI, queue both SET commands, then EXEC to apply atomically -> Option B

4. Quick Check:

   Atomic multi-key update = MULTI + EXEC [OK]

Hint: MULTI + EXEC ensures all-or-nothing updates [OK]

Common Mistakes:

• Running SET commands separately risking partial updates
• Using WATCH without MULTI for atomicity
• Relying on GET before SET without transaction