I’m a big fan of Airflow now, and can’t wait for my next opportunity to use it. The user interface is just – delicious. Airflow orchestrates Python scripts. This project was copying data nightly from SQL Server into Snowflake.
This post started out as “Airflow Integration” and I was going to start with setting up Airflow and dbt, followed by a breakdown of the Python code into sections, before ending with alerts. And maybe I’ll still do that post, sometime, but when I started typing I realised my main “take-away” was… A.I.
Sure I’ve used Python before, but this time seemed different. The code feels 100% mine – which is odd – as I didn’t type any of it, and I didn’t paste it from somewhere else.
“Copilot” is absolutely the right relationship to maintain with the amalgamation engine. As soon as I started asking open questions like “well… what do you think we should do?” it was time to call it a day 🙂
My development experience helped. I found knowing how to get past brick walls, the necessity of frequent testing, when to be pragmatic instead of perfect, and when to go back to the last working version was really helpful.
Ultimately, I feel vibe coding shines for rapid prototyping or as an easy entry point for non-coders. Although debugging was painful, it remains a truly fascinating new tool.
Database Consultant 