26 - Transactions & ACID
511 words
3 min read
Visual companion
Python
Type and operator map
Python Week 1: the first filter for runtime behavior
View
Revision summary
What this note is really saying
Short form
# 26 - Transactions & ACID ## 🎯 Learning Objectives After reading this topic, you will be able to: - Define ACID properties - Describe the transaction state machine - Explain why atomicity and durability are critical - Understand the role of the transaction manager ## 📖 Core Content ### 26.1 Intuition: Why Transac...
26 - Transactions & ACID
🎯 Learning Objectives
After reading this topic, you will be able to:
- Define ACID properties
- Describe the transaction state machine
- Explain why atomicity and durability are critical
- Understand the role of the transaction manager
📖 Core Content
26.1 Intuition: Why Transactions?
Consider a bank transfer: ₹500 from Account A → Account B.
python# Step 1: Read A (balance = 1000) # Step 2: A = A - 500 (balance = 500) # Step 3: Read B (balance = 200) # Step 4: B = B + 500 (balance = 700) # Step 5: Commit
What if the system crashes after Step 2 but before Step 4? ₹500 disappears!
A transaction is a collection of operations that performs a single logical function. The DBMS guarantees that either ALL operations complete or NONE do.
Why This Matters: Without transactions, data integrity is impossible in a concurrent, failure-prone world.
26.2 ACID Properties
| Property | Meaning | How It's Achieved |
|---|---|---|
| Atomicity | All-or-nothing execution | Recovery system (undo log) |
| Consistency | Valid state → valid state | Application + integrity constraints |
| Isolation | Concurrent → appears serial | Concurrency control (locking) |
| Durability | Committed changes persist | Recovery system (redo log) |
Key insight: Consistency follows from atomicity + isolation + integrity constraints.
26.3 Transaction States
(Diagram)
| State | Description |
|---|---|
| Active | Initial state; statements are executing |
| Partially Committed | After final statement, before commit |
| Committed | Transaction completed successfully |
| Failed | Could not complete (error or abort) |
| Aborted | Rolled back to before the transaction |
26.4 Commit and Abort
- Commit: Transaction completed successfully. Changes are made permanent (durability).
- Abort: Transaction failed. Changes are undone (atomicity).
26.5 Schedule
A schedule is the order in which operations from one or more transactions execute.
Serial schedule: Transactions execute one after another (no overlap). Concurrent schedule: Operations from multiple transactions interleave.
A correct concurrent schedule must be equivalent to some serial schedule.
📐 Key Concepts
| Term | Definition |
|---|---|
| Transaction | Logical unit of work with ACID properties |
| Schedule | Order of operations from one or more transactions |
| Serial schedule | Transactions execute sequentially |
| Concurrent schedule | Operations interleave |
⚠️ Common Pitfall
The Mistake: Thinking consistency is entirely the DBMS's responsibility.
Why: The DBMS ensures atomicity + isolation + integrity constraints, but the application must ensure that transactions make sense semantically (e.g., a transfer should debit and credit the same amount).
📝 Practice Questions
Q1. List and explain the ACID properties.
Answer
- Atomicity: Transaction completes fully or not at all
- Consistency: Database moves from one valid state to another
- Isolation: Concurrent transactions don't interfere
- Durability: Committed changes persist after failures
Q2. What is the difference between a serial and a concurrent schedule?
AnswerSerial: Transactions execute one at a time (no overlap, simple but slow). Concurrent: Operations interleave (faster, but must ensure correctness via serializability).
Q3. Draw the transaction state diagram.
AnswerActive → (complete) → Partially Committed → (commit) → Committed Active → (fail) → Failed → (rollback) → Aborted
🔗 Cross-References
- Next Topic: 27 - Serializability
- Related: 30 - Log-Based Recovery (atomicity and durability)
- Textbook: Chapter 14 (Transactions) Join Discord Previous25 - HashingNext27 - Serializability