-- Shape: GROUP BY aggregate on JOIN
-- Join: stackoverflow_posts → comments
-- Description: Grouped aggregate (COUNT, SUM) with GROUP BY on a low-cardinality
-- dimension (post_type_id). Sorted by sub-aggregation metric (SUM). Exercises the
-- DataFusion backend's grouped aggregate pipeline including custom_scan_tlist for scanrelid=0.

-- Query Info (statistics from 100k dataset; larger datasets may have different values):
-- - 'code' selectivity on stackoverflow_posts.body: ~75%

-- Postgres default plan (custom scan off)
SET work_mem TO '8GB'; SET paradedb.enable_aggregate_custom_scan TO off; SELECT p.post_type_id, COUNT(*), SUM(c.score)
FROM stackoverflow_posts p
JOIN comments c ON p.id = c.post_id
WHERE p.body ||| 'code'
GROUP BY p.post_type_id
ORDER BY SUM(c.score) DESC;

-- DataFusion aggregate scan
SET work_mem TO '8GB'; SET paradedb.enable_aggregate_custom_scan TO on; SELECT p.post_type_id, COUNT(*), SUM(c.score)
FROM stackoverflow_posts p
JOIN comments c ON p.id = c.post_id
WHERE p.body ||| 'code'
GROUP BY p.post_type_id
ORDER BY SUM(c.score) DESC;