henry_collins
New Contributor III

It doesn't sound like the === operator itself is the issue, since it should compile into a normal Catalyst expression rather than a SQL function call. The fact that Spark is trying to resolve = as a routine makes me suspect a catalog or analyzer behavior change in DBR 16.4. check whether a custom catalog, Unity Catalog setting, or Spark extension is intercepting analysis before assuming it's a regression. If you can share a minimal reproducible example and your catalog/session configuration, it'll be much easier to narrow down.

Helping users to solve Tech, Software, and Hardware Issues | Web Dev • SEO • Digital Solutions