Question

A data engineer manages a Lakehouse Delta table that receives frequent small incremental MERGE loads throughout the day. Over several weeks, analysts report that queries against the table have become progressively slower, and OneLake storage consumption keeps growing even though the number of active rows is stable. Investigation shows the table's underlying folder contains thousands of small Parquet files plus many obsolete data files from prior versions. Which combination of maintenance operations should the engineer schedule to restore query performance and reclaim storage?

Accepted Answer

Run OPTIMIZE to compact small files into larger ones, then run VACUUM to remove stale data files older than the retention threshold

Answer

Run VACUUM with a retention of 0 hours immediately after every MERGE to delete all historical files

Answer

Recreate the table with CREATE TABLE AS SELECT after each daily load to consolidate files

Answer

Enable V-Order only and rely on it to automatically merge small files and purge old versions