Skip to content

Databricks

Databricks migration from Oracle, SQL Server, MySQL, MariaDB

Last updated:

Datalobe migrates Oracle Database, SQL Server, MySQL and MariaDB to Databricks: it assesses the source, compiles and checks every converted object on Databricks before deploy, loads Delta tables in resumable batches with change data capture (CDC) until cut-over, and validates every table with 17 checks.

Oracle Database / SQL Server / MySQL / MariaDB to Databricks

Source

  • Oracle Database
  • SQL Server
  • MySQL
  • MariaDB

Target

  • Databricks
Code
Databricks SQL

Conversion

What converts to Databricks?

Delta tables are the default table format on Databricks, and Unity Catalog is where Databricks keeps tables, views and permissions.

Source objects and what they become on Databricks
Source objectDatabricks result
Procedures and packagesDatabricks SQL
ViewsViews in Unity Catalog
Tables and dataDelta tables in Unity Catalog

AI converts stored code, every AI conversion is checked, and anything that cannot be proven goes to a named person.

Every converted statement is compiled and checked on Databricks before deploy.

Example run

CRM-Oracle to Databricks-Lakehouse in numbers

In the example run, Oracle Database schema CRM moved to Unity Catalog main.crm on 1 Oct 2026.

Tables
420
Rows
1,604,227,118
Verdict
Passed with warnings

Validation

How is the migration validated?

17 checks per table in six groups show whether the Delta tables match the source.

Questions

Frequently asked questions

Where does the data land in Databricks?

In Delta tables in Unity Catalog. In the example run, the target was the catalogue and schema main.crm.

Is SQL Server supported as a source for Databricks?

Yes. Oracle Database, SQL Server, MySQL and MariaDB are all supported sources for Databricks.

Can we keep using the source during the migration?

Yes. CDC keeps the Delta tables in sync with the source until you cut over. With keep-in-sync (CDC), downtime is limited to the cut-over window.

How are numeric precision differences handled?

As a decision before the run, then checked by validation. In the example run, Oracle NUMBER without precision became DECIMAL(38,10), and Sarah Mitchell accepted 3 warnings.

Next step

Plan your Databricks migration

Start with an Assessment Report on your source database.