A city wants lower cost and faster parking-citation dashboards by combining compute and Delta maintenance improvements. Which pairing fits?

A city seeking lower cost and faster parking-citation dashboards by combining compute and Delta maintenance improvements should compare cost and performance before and after enabling Photon and running OPTIMIZE on the citation tables. Photon accelerates eligible Spark SQL and DataFrame workloads while OPTIMIZE reduces small-file overhead on citation Delta tables; measuring both cost and latency before and after validates whether the pairing actually helps municipal dashboards. Comparing cost and performance before/after enabling Photon and running VACUUM on the citation tables fails because VACUUM only reclaims storage from old snapshots and does not compact small files or improve dashboard query performance, so the before/after comparison won't show the maintenance gain the dashboards need. Comparing cost and performance before/after enabling Photon and running OPTIMIZE on small dimension tables fails because the citation fact tables—not small dimension tables—are where dashboard queries incur cost and latency, so optimizing dimensions leaves the real small-file problem untouched. Comparing cost and performance before/after enabling Photon and running ANALYZE TABLE on the citation tables fails because ANALYZE TABLE only updates query-planner statistics and does not rewrite data files or reduce small-file overhead, so it cannot provide the Delta maintenance gains the dashboards need. Exam caveat: DP-750 optimization experiments should combine Photon and OPTIMIZE with measured before/after evidence. Operational check: capture citation dashboard latency and DBU cost baselines, enable Photon, OPTIMIZE citation tables, remeasure the same queries, and keep or roll back based on evidence.