[Bug] full refresh of failed model with new partition schema fails on __dbt_tmp

[martenlindblad]

Is this a new bug in dbt-bigquery?

• I believe this is a new bug in dbt-bigquery
• I have searched the existing issues, and I could not find an existing issue for this bug

Current Behavior

If I change partition columns of an incremental model and forget to manually do a full refresh the dbt job will throw an error. This is expected. I will then do a full refresh of the model using dbt build --select <model> --full-refresh --target prod --debug and the model is marked as success.
What is not apparent is that the __dbt_tmp table will linger until the next time the full build is run.
The dbt build job will fail on the next run, however, due to the fact that bigquery doesn't allow partition changes in a create or replace statement.

Expected Behavior

A full refresh of a model will drop the __dbt_tmp table.

Steps To Reproduce

1. Change partition schema on incremental model
2. Allow model to fail in dbt build job
3. Do a manual --full-refresh
4. Run the dbt build job again

Relevant log output

12:52:35    Database Error in model <model> (models/<model>.sql)
  Cannot replace a table with a different partitioning spec. Instead, DROP the table, and then recreate it. New partitioning spec is interval(type:day,field:p_date) and existing spec is none
12:52:35  

Environment

- OS: Alpine (container python:3.11-slim)
- Python: 3.11
- dbt-core: 1.8.4
- dbt-bigquery: 1.8.2

Additional Context

No response