3 ms·
Hey everyone, Tobias here, the author. I am a developer fascinated by databases for over a decade and I have consumed (and still consume) everything I can find
by tpetry 3y ago
Hey everyone, Tobias here, the author. I am a developer fascinated by databases for over a decade and I have consumed (and still consume) everything I can find about them - including every article shared here.
When helping other developers, I see that they all struggle with creating correct database indexes for slow queries. But I could never recommend articles or books to learn it as they only discuss very simple examples or focus on super complicated stuff that is only needed by a tiny fraction of developers.
So, I wrote this book to have something I can recommend to everyone. And I did it in a completely different way:
- Focus on only practical stuff a developer needs
- Explanation of the basic principles of how indexes are used
- Examples of how all SQL operations use indexes and how they behave when combined (that's always missed in literature)
- Dozens of visualizations for an easier understanding and to replace boring walls of text
You can download preview chapters on the landing page to experience the difference. If you buy and don't like it, there's a full money-back guarantee with no questions asked.
Let me know if you have any questions, and any feedback is much appreciated!
- XCSme 3y agoThanks, Tobias, bought a copy, maybe I learn something that I can apply to improve the performance of my analytics app schema or indexes [0]. Based on your experience, if the proper indexes are in place, would a ~10m rows MySQL database have similar read performance to a larger, ~100m rows database, if we would aggregate over the most recent ~1m rows in both cases? I am struggling a bit to maintain sub-second queries for databases over ~20m rows, even when the proper indexes seem to be in place. [0]: https://docs.uxwizz.com/guides/database-querying https://docs.uxwizz.com/guides/database-querying