Skip to main content

Databricks Writer

Databricks Writer writes data to Delta Lake tables in Databricks on AWS, Azure, or Google Cloud. Delta Lake is an open-source tabular storage framework that includes a transaction log to support features typically associated with relational databases, such as ACID transactions and optimistic concurrency control.

Databricks Writer can also write to Apache Iceberg tables in Databricks — see Writing to Apache Iceberg Tables.

You can use Striim's Databricks Writer to write data from transactional databases such as Oracle and SQL Server, applications such as Salesforce and ServiceNow, NoSQL databases such as Cosmos DB and MongoDB, object stores such as Amazon S3 and Google Cloud Storage, and other supported sources to Delta Lake tables in Databricks on any of the three major cloud platforms.

Summary of Databricks features common to all cloud providers

Feature

Description

Supported sources

Databricks Writer accepts data from all sources supported by Striim.

Supported write modes

Databricks Writer supports two write modes, Append Only and Merge (with an optional Optimized Merge setting within Merge mode). See Key concepts for Databricks Writer for details on when to use each.

Additional writing features

Supports auto-quiesce after an initial load from Cosmos DB Reader, Database Reader, Mongo Cosmos DB Reader, or MongoDB Reader.

Supports schema evolution to detect and propagate DDL changes from supported sources to the Delta Lake tables.

Supports writing to Apache Iceberg tables in Databricks (Databricks Managed Iceberg and UniForm), in addition to Delta Lake tables. See Writing to Apache Iceberg Tables.

Resilience and recovery

Supports connection retry to avoid application halting due to transient connection issues.

Supports recovery with at-least-once processing.

Performance

Parallel threads can increase throughput to the target in certain situations (not supported when Mode is Merge).

Programmability

Flow Designer, TQL, and wizards in the web UI. See Building pipelines with Databricks Writer.

Metrics and auditing

Key metrics are available through Striim's monitoring features.

Databricks client library

Databricks JDBC driver 2.7.1.

Key limitations

Data is written in batch mode. Streaming mode is not supported in this release. See Databricks Writer operational considerations for additional constraints.

Configuration by cloud endpoint

Use this table to find the attributes that are specific to the cloud platform your Databricks instance runs on.

Attribute

Azure Databricks

Databricks on AWS

Databricks on Google Cloud

Authentication methods

  • Manual OAuth

  • Microsoft Entra ID (formerly Azure Active Directory)

  • Personal Access Token

  • Service Principal (M2M OAuth)

  • Manual OAuth

  • Personal Access Token

  • Service Principal (M2M OAuth)

  • Manual OAuth

  • Personal Access Token

  • Service Principal (M2M OAuth)

Supported staging areas

  • Azure Data Lake Storage (ADLS) Gen2

  • Databricks File System (DBFS)

  • Amazon S3

  • Databricks File System (DBFS)

  • Google Cloud Storage (GCS)

  • Databricks File System (DBFS)

Staging area client libraries

  • azure-identity 1.5.3 (Microsoft Entra ID / ADLS Gen2 authentication)

  • azure-storage-blob 12.18.0 (ADLS Gen2 staging)

  • aws-java-sdk-s3 1.12.589

  • aws-java-sdk-sts 1.11.320

  • google-cloud-storage 2.43.1

Unity Catalog storage credential prerequisite

Not applicable

Not applicable

Requires a Storage Credential (GCP Service Account type) created in Unity Catalog

Note

The Service Principal (M2M OAuth) authentication method works identically regardless of which cloud your Databricks instance runs on. See Key concepts for Databricks Writer and Initial setup for Databricks Writer for the (cloud-agnostic) setup steps.

Note

Databricks File System (DBFS) is available as a staging area on all three clouds, but its use for production or sensitive data is discouraged. See What setup involves.

Note

Unity Catalog storage credential prerequisite: this requirement currently applies only to Databricks on Google Cloud (GCS staging) — see GCS Staging Setup. It does not apply to Azure or AWS staging

For more information

For Databricks on AWS:

For Azure Databricks:

For Databricks on Google Cloud: