Multi-Facet Window Aggregates on JOIN

  • Join: stackoverflow_posts -> comments
  • Description: Uses pdb.agg(...) OVER () window functions to retrieve Top K hits alongside global facet counts across multiple dimensions.

Query Info (statistics from 100k dataset; larger datasets may have different values):

  • ‘code’ selectivity on stackoverflow_posts.body: ~75%

Comparison Strategy

pdb.agg(...) OVER () over a join is only supported by ParadeDB’s Join Scan, so native PostgreSQL (postgres.sql) and Base Scan (basescan.sql) variants are not applicable.