cancel
Showing results forย 
Search instead forย 
Did you mean:ย 
Get Started Discussions
Start your journey with Databricks by joining discussions on getting started guides, tutorials, and introductory topics. Connect with beginners and experts alike to kickstart your Databricks experience.
cancel
Showing results forย 
Search instead forย 
Did you mean:ย 

Tips/Tools/Recommendations for a Snowflake to Databricks Migration?

Office_Solution
New Contributor III

Hey everyone,

We are starting to plan a move from Snowflake over to Databricks and trying to figure out the cleanest way to handle the transition without drowning in manual rewrites.

For anyone who has gone through this:
1) Did you use any parsers or migration tools to convert Snowflake SQL, stored procedures and tasks into Spark SQL/PySpark or just script it yourselves?

2) What tipped you up the most - Handling semi-structured VARIENT data, translating Snowflake specific functions or rethinking compute/clustering?

Would appreciate any practical tips, tool ideas or lessons learned!

1 ACCEPTED SOLUTION

Accepted Solutions

Abhishek20
New Contributor II

Manually rewriting hundreds of Snowflake SQL scripts, stored procedures and tasks into Databricks is a huge time sink.

To skip th grunt work, we used "Pulse Convert" to parse the Snowflake code and automate the bulk translation of standard SQL, views & DDL into Spark SQL and Delta Lake. It knocked out about 75% of the repetitive syntax changes right away so we could focus on the trickier parts. 

Biggest headaches were semi-structured VARIENT parsing, Snowflake-specific functions QUALIFY of FLATTEN edge cases, and shifting away from Snowflake's micro-partitions to Delta Lake liquid clustering.

View solution in original post

2 REPLIES 2

Abhishek20
New Contributor II

Manually rewriting hundreds of Snowflake SQL scripts, stored procedures and tasks into Databricks is a huge time sink.

To skip th grunt work, we used "Pulse Convert" to parse the Snowflake code and automate the bulk translation of standard SQL, views & DDL into Spark SQL and Delta Lake. It knocked out about 75% of the repetitive syntax changes right away so we could focus on the trickier parts. 

Biggest headaches were semi-structured VARIENT parsing, Snowflake-specific functions QUALIFY of FLATTEN edge cases, and shifting away from Snowflake's micro-partitions to Delta Lake liquid clustering.

stephen4
Contributor

One thing Iโ€™d recommend is treating this as more than a SQL syntax conversion. Automated tools can speed up the initial migration, but stored procedures, task dependencies, semi-structured VARIANT data, and Snowflake-specific functions still need careful validation. Iโ€™d start with a representative workload, compare query results between both platforms, and benchmark performance before migrating everything. Also, pay close attention to how your existing clustering strategy translates to Delta Lake. A phased migration with proper testing should help avoid surprises.