In my experience writing software with generic SQL support the example shown here highlights something that's endemic in MySQL. MySQL has a bug (eg. #199) or s

In my experience writing software with generic SQL support the example shown here highlights something that's endemic in MySQL. MySQL has a bug (eg. #199) or some deficiency (no subselects, no transactions, incorrect utf-8 support)[1]. The community's solution is a workaround (Here's how to restructure your select so that you don't need a subselect. Maybe you could use UUIDs instead of integer keys). Meanwhile the problem never existed in Postgres, which seemed to have been correctly and completely implemented from the start. After several years the issue is fixed in MySQL and the boosters ask that the score be reset and the clock re-started. Postgres has its deficiencies and bugs but other than replication they never seemed core to the product in the way that they were for MySQL. [1] These deficiencies span several decades and might seem unfair to bring up but I want to show that what the article is highlighting is just another round of the same problem.