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.
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 object | Databricks result |
|---|---|
| Procedures and packages | Databricks SQL |
| Views | Views in Unity Catalog |
| Tables and data | Delta 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.
Sources (7)
- Databricks documentation: What is Delta Lake in Databricks?, checked
- Databricks documentation: Databricks tables concepts, checked
- Databricks documentation: Unity Catalog securable objects reference, checked
- Databricks SQL language reference: Privileges and securable objects in Unity Catalog, checked
- Databricks SQL language reference: CREATE PROCEDURE, checked
- Databricks SQL language reference: DECIMAL type, checked
- Oracle Database 19c SQL Language Reference: Data Types, checked
Next step
Plan your Databricks migration
Start with an Assessment Report on your source database.