Real APIs return data in wildly different shapes. The skill of ETL (Extract, Transform, Load) is writing logic that pulls the fields you need out of whatever structure you are given. Each dataset here is served in four shapes so you can practice all of them.
| shape | structure |
|---|---|
| flat | A plain array of records: [ {a,b,c}, ... ] |
| list-of-lists | Columns + rows arrays (CSV-like) — forces index-based parsing |
| nested | An object keyed by group: { deptA: { ccX: [ ... ] } } |
| nested-list | Arrays of groups: [ { department, items: [ ... ] } ] |
curl "https://api.ifsjaipur.cloud/playground/data/fpa/actuals?shape=nested" -H "x-api-key: YOUR_API_KEY"Two packs ship today — FP&A (departments, cost centers, monthly budget vs actual) and Stock Market (tickers, OHLCV candles, holdings). The platform is built so more packs (Marketing, HR, Sales) can be added later.
- Discover everything: GET https://api.ifsjaipur.cloud/playground/data lists packs, entities, shapes, and pagination styles.
- In n8n, the "Split Out" node flattens an array field — pair it with shape=flat or nested.
Tip · Practice the same report against all four shapes. Writing the extraction four ways is what makes the skill stick.