by Toussaint Webb, Gopal Lal and Kaitlin Baumgardner

Modern workflows depend on fast, reliable connectivity to data. Whether you’re refreshing dashboards, analyzing data in spreadsheets, or powering applications, the connection layer directly impacts performance and user experience.

As part of our ongoing efforts to improve connectivity to Databricks, we’re excited to share enhancements to the Databricks open-source JDBC driver. Releases 3.x and above introduce significant improvements for partners and customers compared to the legacy 2.x driver:

• Better performance: Delivers up to 30% faster large result retrieval compared to the legacy JDBC driver.
• Improved architecture: Enables new capabilities such as Arrow support for JDK 16+, asynchronous statement execution, and streaming-based volume ingestion.
• New SQL features: Added support for UC metric views, stored procedures, multi-statement transactions, and query tags.  
• Enhanced observability: Built-in client telemetry captures query latency, connection events, and errors, enabling faster root-cause analysis.
• Future proof connectivity: Open-source and fully owned by Databricks, enabling faster fixes, code transparency, external code contributions, and tighter platform integration.

When Databricks released its OSS JDBC driver last year, the migration was seamless for us. We were able to maintain backward compatibility while gaining faster access to new features, capabilities, and fixes. That has helped us shorten time to market and bring support for new Databricks innovations, including UC Business Semantics, to customers more quickly. —Jamie Davidson, President & Co-founder, Omni

Better performance where it matters most

For many BI and application workloads, retrieving large datasets is the biggest performance bottleneck. The OSS JDBC driver significantly improves performance for these scenarios.

When returning large query results, the new driver delivers up to 30% faster performance compared to the legacy JDBC driver. 

These improvements are especially impactful for organizations running operational analytics or high-volume reporting workloads on Databricks.

Improved Architecture

The new Databricks JDBC driver has seen improvements in the underlying architecture.

• Arrow compatibility for JDK 16+: Supports full Arrow-based data transfer on modern JVMs without workarounds, allowing customers and partners to keep Arrow enabled and benefit from its performance gains. 
• Asynchronous execution interface: Extends JDBC with a first-class async API, so applications can submit queries and keep working while results are computed, enabling more responsive architectures and better resource utilization.
• Stream-based Volume ingestion: Streams bulk data directly into Databricks Volumes without local staging, removing disk I/O bottlenecks and making large ingestion workflows faster and easier across apps, pipelines, and ETL tools.
• Statement Execution API: Integrates with Databricks’ Statement Execution API to enable direct, programmatic query execution with improved control over execution lifecycle, making it easier to build responsive applications and automate workflows.

Expanded SQL capabilities for modern applications

The new Databricks JDBC driver also introduces new functionality that enables richer database-style workflows and more sophisticated integrations.

New capabilities include support for:

• Stored procedures, making it easier to encapsulate business logic and simplify application development
• Multi-statement transactions, enabling more complex workflows with transactional guarantees
• Unity Catalog metric views, enabling customers to seamlessly interact with their semantic layer in third-party tools
• Query Tags, enabling users to label and track queries for improved observability, cost attribution, and workload management
• Geospatial data type, enabling native storage and analysis of location-based data for richer spatial insights and use cases
• Complex data types, enabling native handling of maps, arrays, and structs with familiar Java-style semantics for more flexible data modeling and processing

These features help teams build better applications that take full advantage of the latest innovations in Databricks.

Better observability 

The new Databricks OSS JDBC driver ships with built-in client telemetry that captures near real time query latency, metrics and errors, without affecting query performance. For customers and partners, this translates to faster turnaround on support cases, more precise fixes, and a driver that gets measurably better over time as real-world usage patterns inform every release.

A more future-proof connectivity layer

One of the biggest long-term benefits of this release is that Databricks owns and maintains the JDBC driver codebase. Compared to the legacy JDBC driver, this means:

• Faster bug fixes 
• Faster delivery of new features
• Closer alignment with platform capabilities
• Open-source code transparency and community contributions 

This translates into a connectivity layer that evolves at the same pace as the Databricks platform itself.

Getting started

The open-source Databricks JDBC driver marks an important step forward for connectivity to Databricks. With an improved architecture, faster performance, expanded SQL capabilities, and deeper platform integration, you can build more reliable data experiences on Databricks.

To see the full list of recent updates, review the latest release notes, access the driver through Maven, or try the new driver in your environment today.