fix: use explicit seed types or null type casts where necessary for postgres

OHDSI / dbt-synthea

[Under development] A dbt ETL project to convert a Synthea synthetic data set into the OMOP CDM

https://ohdsi.github.io/dbt-synthea/

Apache License 2.0

16 stars 6 forks source link

fix: use explicit seed types or null type casts where necessary for postgres #70

Closed lawrenceadams closed 1 month ago

lawrenceadams commented 1 month ago

This PR fixes a handful of issues that cause postgres to fail running the project.

These issues were all around type casting, given the empty seed files would result to everything being cast as integrers, and nulls without types being treated as text types. Duckdb is much more forgiving!

For now I have:

Added explicit seed types as needed (only on files where the seed is blank)
Explicitly cast NULLs to INTEGER types for ids

lawrenceadams commented 1 month ago

Good catch! I hadn't run dbt test on postgres recently so I missed these 😅

Neither had I haha 🙃

lawrenceadams commented 1 month ago

@katy-sadowski I have applied the dbt.cast here - I think it's best to clean up all other casts in another PR or this will get messy!

edit: passes on pg.