A useful way to iterate on queries is to use the TABLESAMPLE feature to speed up processing time. However, this takes a ~random sample of the dataset and is not suitable for production use.
Add a linter rule to catch any sampling that might slip into analysis PRs.