Count distincts are one of the most expensive operations a database can do. Luckily, approximate count distinct works just as well for most use cases. If someone is using count distinct instead of the approximate, reject the query and inform them of the alternative. (Or alternatively, replace count distincts automatically with approximate count distinct.)
We've all seen the brutal SELECT * from MASSIVE_FACT_TABLE query. No predicate nor selective set of columns. Just a lot of work for little purpose. With Sundeck, use flows to automatically reject queries against large tables that aren't filtered appropriately.
It's true in our society and also in databases. Inequality joins can make operations far more expensive. Sundeck can watch for inequality joins and makes sure the right users can execute them but others cannot.
While a lot of joins only require a single condition, there are also those tables where multiple conditions are required and are not immediately obvious. Forgetting can hurt everyone--best case, it results in extra resource consumption. Worse case, users make incorrect conclusions from data. Why not just nip that in the bud? Just tell Sundeck the specific join conditions expected for table pairings, and reject the query if those conditions don't exist.
In a few quick steps, you can start using Sundeck to enhance your Snowflake environment.
Get more productive today!
