Databricks SQL
activeby Databricks · since 2020 · also known as DBSQL, Databricks SQL Warehouse, Databricks SQL Serverless
Serverless SQL warehouses on the Databricks lakehouse, powered by the Photon C++ engine.
Overview
Databricks SQL is the warehousing layer of the Databricks platform: SQL warehouses (classic, pro or serverless) run the vectorized Photon engine directly over Delta Lake and Iceberg tables governed by Unity Catalog. Because the data stays in open formats in your cloud storage, the same tables serve Spark jobs, ML and BI without copies. It adds materialized views, streaming tables, AI functions and predictive optimization. It is strongest for organizations already on Databricks that want to retire a separate warehouse; for pure SQL shops it can feel like more platform than they need.
Strengths & trade-offs
Strengths
- +Warehouse over open Delta Lake/Iceberg tables in your own cloud storage
- +Photon vectorized engine gives strong price/performance on large scans
- +Unity Catalog provides one governance layer for SQL, ML and AI assets
- +Serverless warehouses start in seconds and scale for concurrency
- +Streaming tables and materialized views for incremental pipelines
Trade-offs
- −DBU pricing plus underlying cloud costs is hard to forecast
- −Platform breadth means a steeper learning curve for SQL-only teams
- −Small, frequent writes create small-file and optimization overhead
- −No on-premises option
Great for
- +Lakehouse SQL and BI on Delta Lake
- +Consolidating a separate warehouse into Databricks
- +dbt transformations on the lakehouse
- +Serving features and aggregates for ML
Think twice for
- −OLTP workloads
- −Sub-second user-facing analytics at high QPS
- −Small teams with only a few GB of data
Capabilities
Native or official first-party support · 15 of 25 tracked
Scale & consistency
- Comfortable data size
- Massive · Multi-PB
- Horizontal scaling
- Native
- Multi-region
- No
- Transactions
- Full ACID
- Consistency
- Strong
- Schema
- Flexible
- Architecture
- Serverless, Distributed, Shared storage
- Query languages
- SQL (Databricks SQL / ANSI), Python
Warehouses autoscale clusters for concurrency; storage is your own object storage, so capacity is effectively unbounded.
Deployment
available on · AWS · Azure · Google Cloud
Ecosystem
Alternatives to Databricks SQL
Fully managed cloud data warehouse with separate storage and compute, now spanning AI and apps.
databricks-sql vs snowflake →Google's serverless, pay-per-query data warehouse built on Dremel and Colossus.
databricks-sql vs google-bigquery →T-SQL warehouse in Microsoft Fabric that stores every table as Delta Parquet in OneLake.
databricks-sql vs microsoft-fabric-warehouse →Commercial Trino platform: self-managed Starburst Enterprise and the Starburst Galaxy SaaS.
databricks-sql vs starburst →Iceberg-native lakehouse query engine with reflections for acceleration, owned by SAP since 2026.
databricks-sql vs dremio →Spark's SQL and DataFrame engine for large-scale batch, streaming and lakehouse processing.
databricks-sql vs apache-spark-sql →Databricks SQL FAQ
What is Databricks SQL?
Databricks SQL is a columnar / OLAP database by Databricks, first released in 2020. Databricks SQL is the warehousing layer of the Databricks platform: SQL warehouses (classic, pro or serverless) run the vectorized Photon engine directly over Delta Lake and Iceberg tables governed by Unity Catalog. Because the data stays in open formats in your cloud storage, the same tables serve Spark jobs, ML and BI without copies. It adds materialized views, streaming tables, AI functions and predictive optimization. It is strongest for organizations already on Databricks that want to retire a separate warehouse; for pure SQL shops it can feel like more platform than they need.
Is Databricks SQL open source?
No. Databricks SQL is proprietary, closed-source software.
Does Databricks SQL support ACID transactions?
Yes. Databricks SQL supports full ACID transactions, with strong consistency.
Does Databricks SQL scale horizontally?
Databricks SQL scales horizontally out of the box, with built-in sharding or partitioning. It is comfortable at massive scale (Multi-PB).
How can I run Databricks SQL?
You can use the vendor's managed cloud and run it serverless. Managed offerings include Azure Databricks. It is available on AWS, Azure and Google Cloud.
What is Databricks SQL best for?
Databricks SQL is a strong fit for lakehouse SQL and BI on Delta Lake; consolidating a separate warehouse into Databricks; and dbt transformations on the lakehouse. Think twice for OLTP workloads.
What are the best alternatives to Databricks SQL?
Common alternatives to Databricks SQL include Snowflake, Google BigQuery, Microsoft Fabric Data Warehouse, Starburst and Dremio. The right choice depends on your workload, scale and hosting; WhichDB's smart match ranks them against your requirements.
How popular is Databricks SQL?
Databricks SQL ranks #147 of 844 databases on the WhichDB Score (20.8/100), and #15 among columnar / OLAP databases. The score combines public adoption signals such as package downloads, GitHub activity, job posts and developer discussion.