Why should you care?

Flipkart product search on PostgresVerse specs jsonb needs GIN — B-Tree cannot index “contains key warranty inside JSON”.

See it live — copy this example

Run in pgAdmin or psql.

CREATE INDEX idx_products_specs_gin
ON products USING GIN (specs jsonb_path_ops);

SELECT name, specs
FROM products
WHERE specs @> '{"warranty":"2y"}';

What happened?

• jsonb_path_ops GIN index supports @> containment.
• Query finds products whose specs include warranty 2y without scanning every json blob.

Practice next

1. Ensure products.specs is jsonb with varied documents.
2. CREATE INDEX USING GIN as shown.
3. Run EXPLAIN ANALYZE on the @> query.
4. Index tags text[] with GIN and query WHERE tags @> ARRAY['sale'];
5. Compare jsonb_ops vs jsonb_path_ops index sizes.

Remember

GIN powers jsonb, array, and FTS lookups. @> checks json containment. GIN indexes are larger and slower to build than B-Tree.

PostgresVerse facet filters

Category page filters laptops with 2y warranty via jsonb @> using GIN.

Outcome: Filter API stays under 50ms on 2M SKUs.