What Is New in MongoDB 7.2
MongoDB 7.2 delivers a set of enhancements focused on developer productivity, operational efficiency, and data security. The release introduces new query capabilities, simplifies data encryption, and improves performance for time series collections.
| Category | Key Changes |
|---|---|
| New Features | Vector Search General Availability, Compound Wildcard Indexes, On-Demand Materialized Views |
| Query Enhancements | New Aggregation Operators ($dateAdd, $dateSubtract, $bitAnd, etc.), $unionWith array support |
| Time Series | Delete Operations, Sharding Support, Columnar Compression |
| Security | Queryable Encryption GA, New Key Management Integrations |
| Operability | Balancer Control, Initial Sync Progress Monitoring |
What query operators were added in MongoDB 7.2?
The aggregation framework gets a significant boost with new operators for date and bitwise manipulation. This makes complex data transformations directly in the database more powerful and reduces the need for application-side processing.
New date operators like $dateAdd and $dateSubtract provide precise control over datetime calculations. For bitwise operations, $bitAnd, $bitOr, $bitXor, and $bitNot are now available. The $unionWith stage also gains support for piping in an array of documents, not just a collection name.
How is Vector Search different in GA?
Vector Search has moved from preview to General Availability, signaling its production readiness. This is a big deal for building AI-powered applications like semantic search, recommendation engines, and anomaly detection directly on your operational data.
In practice, this means the feature is fully supported for deployment. You can now create vector search indexes and perform nearest neighbor searches on vector embeddings with confidence, all natively inside MongoDB without a separate vector database.
Can I delete data in time series collections now?
Yes, deletion operations are now fully supported for time series collections. You can use deleteOne(), deleteMany(), and findAndModify() with remove: true to remove documents based on your criteria.
This was a major limitation in earlier versions. Now, managing the lifecycle of time series data"like purging old records for compliance or correcting erroneous measurements"is completely seamless and doesn't require workarounds.
What are compound wildcard indexes?
Compound wildcard indexes allow you to create an index that combines a wildcard pattern on one or more fields with specific indexes on others. This is a powerful extension of the existing wildcard index functionality.
For example, you can now create an index on { "metadata.$**": 1, "userId": 1 }. This is perfect for scenarios where you have a mix of predictable query patterns (like filtering on userId) and unpredictable patterns on dynamic sub-documents.
Is Queryable Encryption ready for production?
Queryable Encryption has also achieved General Availability status in 7.2. This technology allows you to run expressive queries on fully encrypted data without decrypting it server-side, a major step forward for data privacy.
With this release, it's considered production-ready for supported query types. MongoDB has also expanded its integrations with external key management providers like Google Cloud KMS and Azure Key Vault, giving you more flexibility in managing encryption keys.
FAQ
What is the most significant performance improvement in 7.2?
For time series workloads, the introduction of columnar compression is a game-changer. It can significantly reduce storage footprint and improve query performance for large-scale time series data by storing data more efficiently.
Can I shard my time series collections?
Yes, sharding support for time series collections is now available. This allows you to distribute your time series data across a cluster, enabling horizontal scaling for massive IoT and telemetry datasets that outgrow a single server.
How does the balancer control work?
New commands like balancerStart and balancerStop give you more granular programmatic control over the shard balancer. This is useful for automating maintenance windows or coordinating balancer activity with other cluster operations.
What are On-Demand Materialized Views?
This feature lets you create a materialized view that is updated manually via the $merge command, not automatically on source collection changes. It offers a performance boost for complex aggregations when real-time data isn't a strict requirement.
Is it easier to monitor an initial sync now?
Absolutely. The currentOp output and diagnostic logs now provide more detailed progress reporting for initial sync operations. You get better visibility into how long each phase takes and the overall status, which helps with troubleshooting and planning.