A short journey, start to finish: one mandatory setting, then three quick decision points — Delta, Joins, and Scaledown & Shuffle. Answer as you go, and your full config is waiting at the end.
Delta Lake
Joins
Scaledown & Shuffle
Mandatory — always
Compact small files automatically
spark.databricks.delta.autoCompact.enabled = true
Why
Every write in Fabric (batch or streaming) creates new files. Left uncompacted, a table ends up with hundreds of tiny files — every read has to open all of them. That penalizes every consumer: Power BI, the SQL endpoint, other notebooks.
When it applies
Always. This is the only setting on this page with no conditional scenario — the cost of enabling it is minimal and the benefit is universal.
Now walk through the 3 decision points
End of the journey
Your full config
–
Optional active
–
Total optional
Saved only on this browser (localStorage) — nothing is sent anywhere.
