Hi,
@carolus_holman! As
@edgar_ramirez_mondragon points out, we've got a WIP database stream type coming soon. That implementation uses SQLAlchemy behind the scenes (in turn powered by DB API 2.0), with the expectation that developers would only need to override parts of the process for which a custom approach is significantly faster/better than the default SQLAlchemy approach.
Is your use case expected to be high volume? If not, I'd be curious if our generic
tap-sqlalchemy might work for your case - and/or if you could fork it and add Trino-specific optimizations as needed. The SQLAlchemy tap should still be considered pre-release, but you can try it out if you wanted. It should give generic extract ability for any installed sqlalchemy drivers on your system.