Senior Spark interviews always ask how you prove whether Catalyst chose a BroadcastHashJoin or fell back to SortMergeJoin.
dfs['orders'] to valid positive rows (amount > 0).F.broadcast(dfs['regions']) on 'region_id'..explain() so BroadcastHashJoin is verified.order_id, region_name, amount sorted by order_id.These DataFrames are passed to solve(spark, dfs). Access one with dfs['name'].