Jeff Klukas

#9 should hopefully be resolved now. Are you proposing that we support `CreateView(..., or_replace=True)` for SQLite by rewriting to "DROP VIEW IF EXISTS" and "CREATE"?

I assume this module generally works with SQLite, but throws an error if you set `or_replace=True`. Is that correct?

My general impression is that materialized views are a fairly implementation-specific concept with significantly different syntax per db. It may be more reasonable for individual dialects to implement appropriate constructs...

I haven't personally used the ORM much, so the use cases there aren't immediately obvious to me. I'll try to take more of a look at your suggestions later this...

I'm not convinced that we gain much by actually moving the table to a historical location. I'd like to see a way of marking a docType as deprecated, perhaps via...

Note on implementation: this test will likely need to pull the pre-PR state of any changed schemas via some `git` commands, then do some amount of parsing of the two...

> should we move that to a separate issue Indeed. I interpreted this as being about CI tests. Apologies for missing the scope. I agree that adding to a checklist...

Other top-level fields to be added during ingestion that we should disallow: - submission_timestamp - sample_id - Anything prefixed with `normalized_`

The Dockerfile includes `RUN mvn dependency:copy-dependencies` which is how `target/ingestion-sink-0.1-SNAPSHOT.jar` ends up in the built image at `/app/target/ingestion-sink-0.1-SNAPSHOT.jar`. In the `mps-test` script, I see now there a comment: ``` #...

Thanks for writing this up, @ahounsel!