Loading Millions of Rows of Test Data in Seconds
summary
A live demo generates 10 million rows of realistic customer data in about 17–19 seconds using TSQL only, a numbers table doubling trick and modulo name reuse with a 280,000-row NPPES names dump; no third-party tools. It inserts set-based, fast batches, produces parent/child variants, exposes missing indexes and CPU‑heavy queries, and completes in seconds so the performance claim and the built scripts are directly verifiable.
AI-written summary · about
related talks
53mOptimizing your Azure SQL Database workloads
52mExploring PL/SQL New Features and Best Practices for Better Performance
29mHow to build a data pipeline without data
47mEfficiency up, failed releases down; compliant production data for your test and dev environments
48mOptimal Data Loads
23mChallenges of database testing
58mTroubleshooting CPU Performance Issue for SQL Developers
47mLoad Testing Storage Performance for SQL Server