While environments like 'dbt' can bring some reasoning, testability, jinja macros and iterative constructs to SQL, it also can help contain it. At my organization I advocate for the minimization of the SQL surface, and we utilize DRYer, modern, testable languages wherever feasible, like Clojure (mostly) and Python. Clojure integrates very well with Kafka, and it is used throughout our organization. While Java/Clojure lacks some of the data applications found in the Python ecosystem, we certainly use Python where appropriate.