This repo is a fork of dbt-external-tables with a few minor changes to better work with Databricks. It is not intended to be a full replacement for the original project.
Spark & Databricks can automatically detect and recover partitions and infer full schema without
explicitly specifying the schema or partitions in the sources specification YAML files,
but this is not supported by dbt-external-tables version 0.9.0.
This fork allows automatic recovery of the partitions when recover_partitions
flag is added:
- name: my_external_table
external:
using: parquet
location: "<path to files>"
recover_partitions: true
Version 0.9.0 of dbt-external-tables does not work correctly with Databricks Catalogs. This causes stage_external_sources
to always drop tables and do a full refresh, which is very slow. This is fixed here, but probably breaks compatibility with Spark.