How does MongoDB handle transactions?

In MongoDB, the capability for handling transaction data was introduced in version 3.2 with support for atomic single-document operations, and in version 4.0, multi-document transactions were officially introduced, enabling MongoDB to handle transaction data more robustly and flexibly.

1. Multi-Document Transactions

Starting from MongoDB 4.0, it supports multi-document transactions, allowing atomic write operations across multiple documents. This is common in SQL databases but represents a significant advancement in NoSQL databases.

How it works:

• Use startTransaction to begin a new transaction.
• Execute the required read and write operations.
• Use commitTransaction to commit the transaction; if the operation succeeds, all changes are permanently saved.
• If an error occurs within the transaction or changes need to be rolled back, use abortTransaction to revert all changes.

Practical Application Example:

Assume we are managing a database for an e-commerce platform, needing to update inventory and record an order. Both operations must be executed within the same transaction to ensure data consistency and accuracy.

javascript
const session = db.getMongo().startSession();
session.startTransaction();
try {
    db.products.updateOne({ _id: productId }, { $inc: { stock: -1 } }, { session });
    db.orders.insertOne({ productId: productId, quantity: 1, status: 'confirmed' }, { session });
    session.commitTransaction();
} catch (error) {
    session.abortTransaction();
    throw error;
} finally {
    session.endSession();
}

2. Atomic Single-Document Operations

Before the introduction of multi-document transactions, MongoDB already supported atomic operations within a single document, which remains an effective means for ensuring data consistency when handling individual documents.

How it works:

• Use update operators such as $set, $inc, $push, etc., to modify a single document in one operation without requiring a read-before-write step, reducing the risk of data inconsistencies during the operation.

Practical Application Example:

For instance, adding points to a user's account:

javascript
db.users.updateOne({ _id: userId }, { $inc: { points: 100 } });

3. Transactions and Replica Sets

MongoDB transactions rely on the structure of replica sets. All transaction write operations require confirmation from a majority of the replica set members, ensuring high availability and data consistency.

4. Performance Considerations

Although transactions provide strong data consistency guarantees, they may impact performance due to higher resource usage and coordination overhead. Therefore, it is crucial to use transactions judiciously when designing applications to avoid unnecessary overhead.

In summary, MongoDB, by introducing multi-document transactions, provides transaction handling capabilities similar to traditional relational databases while maintaining the flexibility and scalability of NoSQL databases. This enables MongoDB to more effectively handle complex scenarios requiring high data consistency.