Reject a query with a custom defined error message. Helpful for guiding users (e.g. "don't forget the second join condition") and avoiding anti-patterns ("apply a day filter when reading the giant fact table"). Can also be useful for setting time-of-use rules on certain warehouses or datasets.
Execute a Snowflake statement before executing your main query. This can help to setup certain operations. For example, route to an overflow warehouse if the primary warehouse is overloaded. It can also allow you to build advanced granular quota systems such as "don't let a user execute more than $5/day of queries". Your SQL creativity is your limit.
Syntax and context aware replacement of one or more tables when they occur. This can be helpful in many situations including when you're running test jobs and want to replace large fact tables with small alternatives.