We have had a couple of cases where developers put hard coded table references in models instead of using refs and sources, which caused some issues at a later point. Ideally code review would have flagged these, but some cases of the scenario were missed, and similar cases potentially could be missed in the future. We would ideally like to incorporate some sort of linting or other types of checks as part of our CI/CD process to flag such issues. Wondering if anybody out there has implemented any such checks as a part of CI/CD, or is aware of a dbt plugin that might help with this type of issue.
1 Like
dbt-checkpoint seems to have a check for this, although I have not tried using it myself: https://github.com/dbt-checkpoint/dbt-checkpoint/blob/main/HOOKS.md#check-script-has-no-table-name
Note: @Owen originally posted this reply in Slack. It might not have transferred perfectly.
1 Like