Why should you care?

Flipkart filters PostgresVerse products by brand and RAM without exploding into 50 nullable columns.

See it live — copy this example

Run in pgAdmin or psql.

SELECT name, specs
FROM products
WHERE specs @> '{"brand":"Nova","ram_gb":8}'
  AND (specs->>'in_stock')::boolean = true
ORDER BY (specs->>'price')::numeric;

What happened?

• @> checks nested containment.
• ->> returns text for cast to boolean/numeric.
• Mixed relational columns and jsonb predicates in one query.

Practice next

1. Insert products with varied specs jsonb.
2. Run containment query; add/remove keys in JSON.
3. Try jsonb_path_exists(specs, '$.warranty ? (@ == "2y")').
4. Update one key: specs = jsonb_set(specs, '{color}', '"black"').
5. Aggregate: SELECT specs->>'brand', count(*) GROUP BY 1.

Remember

@> is primary filter for indexed jsonb. ->> returns text; cast to proper type. jsonpath handles complex document logic.

PostgresVerse laptop finder

Shoppers filter 8GB RAM Nova laptops in stock via jsonb API filters.

Outcome: Merchandising adds new spec keys without migrations every week.