Best Practices for Row-Level Deletion in SQLMesh Projects

[SilasFriby]

Hi SQLMesh team,

I'm working on a project using SQLMesh and was wondering what the preferred method is for deleting specific rows (e.g., for a specific customer ID).

We’re building a GDPR-compliant deletion pipeline, and I'm currently considering writing a script that scans all physical tables across environments and deletes rows that match certain criteria.

Is that the right approach? Or is there a recommended “SQLMesh way” to handle row-level deletions?

Thank you in advance!

Best regards,
Silas

[cmgoffena13]

Hey @SilasFriby -- a little late but I am starting to work through this as well. SQLMesh maintains the data in the tables by tracking "intervals," so it doesn't track individual rows -- it does not have any functionality to delete out specific records. The best approach right now would be target the SQLMesh backend schemas, sqlmesh__xxx, and delete the rows out of all of the physical tables.

I'd curious to hear from the community how they're handling GDPR or CCPA compliance within the SQLMesh framework!