Which command treats incremental models as table models?

The command that treats incremental models as table models is the option that includes the --full-refresh flag. When this flag is used during the dbt run command, it forces dbt to drop and recreate the tables instead of performing an incremental update. In typical usage, incremental models allow for data to be added to existing records based on specified criteria (like timestamps or unique identifiers) without reprocessing the entire dataset. However, invoking a full refresh overrides this incremental behavior, ensuring that the model behaves like a standard table model—essentially reloading the entire dataset from scratch.

This feature is particularly useful when there have been significant changes or updates in the source data that an incremental model might not adequately capture. By treating the incremental model as a table, the user ensures the output is fresh and reflective of the most current data state.

Other commands, such as dbt clean, are designed for clearing out target directories or compiled artifacts but do not affect how models are processed. The dbt debug command is used to check the configuration and connection settings without executing model runs. The basic dbt run command, without --full-refresh, operates standardly, respecting the incremental nature of the models.