fix: unrecognized _partitiontime for time ingestion partitioned models

[malikfm]

Description

Update incremental runner to support models with time ingestion partitioning.

Fixes #73

Checklist:

• I have run make verify and fixed any linting or test errors
• I have added appropriate unit tests or if applicable an integration test
• OPTIONAL: I have run make integration against a Big Query instance

[ccharlesgb]

How come we are removing the field from the predicted schema? If time ingestion partitioning is enabled then I can see how _PARTITIONTIME gets added to the predicted schema but I don't see why you would remove the partition_by field from the predicted schema?

[malikfm]

Assuming time_ingestion_partitioning is set to true, I thought that there wouldn't be any queries that select the partition_by.field since it essentially won't be a field in the real table. I put it back to the predicted schema.

[ccharlesgb]

If the incremental already exists does dbt behave differently? I don't think dbt can modify the partition keys for tables once they have been created without a full refresh

[malikfm]

I reverted this line back to the original.

[ccharlesgb]

Seems like the latest python 3.12 version has broken something 🤔 Will look into this

[ccharlesgb]

Should be good to merge once you rebase from master 👍

[malikfm]

Thanks 👍

[ccharlesgb]

This is release in v0.8.0. I think we still might be missing some edge cases here like if the table already exists but isn't partitioned and then someone adds time_ingestion_partitioning later on then dbt won't update the partitioning of the table but the dry run will think it will. But happy for you to raise more PRs if you see this come up.