Maybe you just don't know what you are doing.
Maybe you just have nothing to say and don't recognize that such behavirs are clearly wrong, that failing silently for a database is a case of bad design, and MySQL is just technically inferior to Postgres. Or are we missing something?
Yes you are. The first thing you are missing that such behaviors are configurable: you can choose whether to fail silently or get an error.
The second thing you are missing is the reason such behavior exist in the first place — namely, there are usage scenarios when it is preferable. Having ability to choose which trade-offs to have is in no way technical inferiority.
Same goes for the date. And please don't tell me the error message you get back there is quite correct. That alter table statement tries to silently add a default and later fails due to the default not being valid.
Also, see this:
this table now contains two columns that can't be null and yet the default is null. Thanks.Concerning corrupted tables. I have around 1.5 TB of data spread over ~10 databases and over the last 10 years, I never had a corrupt table that I could not trace back to hardware failure (harddrives, raid controllers, RAM).
Over the same period, I had around 200 MB of data spread over ~10 databases in MySQL and I had three cases of corrupt tables that I could not trace back to defective hardware as various diagnostic tools yielded no error and the machines still run without problems with PostgreSQL.
Maybe I'm just unlucky.