Question

You are building a demo environment and need Copilot to generate 200 rows of realistic seed data for a relational database with three tables: `customers`, `orders` (which references `customers.id` via a foreign key), and `order_items` (which references `orders.id`). Which approach is most likely to produce sample data that actually loads without foreign-key constraint violations?

Accepted Answer

Open the schema/DDL file for all three tables in the editor and prompt Copilot to generate the INSERT statements in dependency order (parents before children), explicitly stating that child rows must reference existing parent IDs.

Answer

Ask Copilot in three separate, unrelated chat sessions to generate each table's rows independently, then concatenate the output into one script.

Answer

Prompt Copilot to generate all 200 rows for each table using fully random UUIDs for every id and foreign-key column so the values look realistic.

Answer

Request only the `customers` and `order_items` data and let the database auto-populate the `orders` table through cascade defaults.