I created an article to show differences in materializations for
- view
- table
- incremental
- snapshot
in combination with a persistent (never gets truncated) and transient source/staging (gets truncated before the next load) table.
The article should help to decide what materialization to use if the source is persistent or transient: https://www.buckenhofer.com/2022/11/materialization-examples-of-data-engineering-with-dbt/
code is on github: GitHub - abuckenhofer/dbt: dbt materialization scenarios