- Home
- Data Version Control Tools
- Delta Lake
Delta Lake
What it is
Delta Lake is a free, open-source storage framework for building format-agnostic lakehouse architectures on data lakes. It supports serializable ACID transactions to protect data consistency, along with schema enforcement and evolution to manage changes to table data. SQL, Scala/Java, and Python APIs support merging, updating, and deleting datasets. Batch and streaming workloads share a system with exactly-once ingestion semantics, while table time travel provides access to earlier versions for audits, rollbacks, or reproducibility.
Its scalable metadata is designed for petabyte-scale tables with billions of partitions and files. Delta Universal Format lets Iceberg and Hudi clients read Delta tables. Listed integrations and compatible engines include Spark, Flink, Hive, Trino, Athena, BigQuery, Databricks, Microsoft Fabric, Snowflake in beta, PrestoDB, and Redshift. Delta Lake can be used across cloud, on-premises, and local environments with compatible compute engines. It uses a bring-your-own storage backend and supports self-hosted deployment. The independent project is governed through the Linux Foundation Projects, and releases are compatible with specified Apache Spark versions.
For and against
For
- Free and open-source storage framework.
- Serializable transactions support data consistency.
- Time travel gives access to earlier table versions.
- Supports batch and streaming workloads.
- Iceberg and Hudi clients can read Delta tables.
Against
- Requires a bring-your-own storage backend.
- Releases are compatible with specified Apache Spark versions.
- Writing to a shared table is not supported by Delta Sharing.
The verdict
Delta Lake provides table storage features for lakehouse architectures, including transactions, schema controls, and earlier-version access. Its bring-your-own storage model and Spark version compatibility are important considerations when planning a deployment.
Who it suits
Delta Lake suits teams building lakehouse architectures that need table transactions, version access, and batch and streaming workloads. It can be used with compatible compute engines across cloud, on-premises, and local environments.
Plans and prices
Delta Lake
Open-source storage framework; use with compatible compute engines and platforms
Read 7 Oct 2026 · delta.io
On the record
- ACID transactions
- Delta Lake provides serializable transactions to protect data consistency.delta.io · read 7 Oct 2026
- Batch and streaming
- Delta Lake supports unified batch and streaming workloads with exactly-once ingestion semantics.delta.io · read 7 Oct 2026
- Community support
- The project points users to Slack, GitHub Issues, Stack Overflow, and Google Groups for community help.delta.io · read 7 Oct 2026
- Compatibility requirement
- Delta Lake releases are compatible with specified Apache Spark versions; for example, the releases page lists Delta Lake 4.3.x with Spark 4.0.x and 4.1.x.docs.delta.io · read 7 Oct 2026
- Compute engines
- The site lists Spark, PrestoDB, Flink, Trino, Hive, Snowflake, Google BigQuery, Athena, Redshift, Databricks, and Azure Fabric among compatible compute engines.delta.io · read 7 Oct 2026
- Data changes
- SQL, Scala/Java, and Python APIs support merge, update, and delete operations on datasets.delta.io · read 7 Oct 2026
- Data operations
- Delta Lake provides SQL, Scala/Java, and Python APIs to merge, update, and delete datasets.delta.io · read 7 Oct 2026
- Governance
- Delta Lake is an independent open-source project governed through the Linux Foundation Projects and is not controlled by a single company.delta.io · read 7 Oct 2026
- Integration example
- The integrations page says Apache Spark can read from and write to Delta Lake.delta.io · read 7 Oct 2026
- Integration limit
- The integrations page describes its ClickHouse integration as read-only for existing Delta Lake tables in Amazon S3.delta.io · read 7 Oct 2026
- Integrations
- Listed integrations include Apache Spark, Flink, Hive, Trino, Athena, BigQuery, Databricks, Microsoft Fabric, and Snowflake (Beta).delta.io · read 7 Oct 2026
- Interoperability
- Delta Universal Format lets Iceberg and Hudi clients read Delta tables.delta.io · read 7 Oct 2026
- Open format
- Delta Universal Format (UniForm) lets Iceberg and Hudi clients read Delta tables.delta.io · read 7 Oct 2026
- Project governance
- Delta Lake is an independent open-source project that the site says is not controlled by any single company and joined the Linux Foundation Projects in 2019.delta.io · read 7 Oct 2026
- Scalable metadata
- The project says it can handle petabyte-scale tables with billions of partitions and files.delta.io · read 7 Oct 2026
- Schema controls
- Schema evolution and enforcement help prevent bad data from causing corruption.delta.io · read 7 Oct 2026
- Sharing limit
- Delta Sharing reads support batch, streaming, and table changes, but writing to a shared table is not supported.docs.delta.io · read 7 Oct 2026
- Table scale
- Its scalable metadata feature is described as handling petabyte-scale tables with billions of partitions and files.delta.io · read 7 Oct 2026