SQL is a unique case. It's important enough that you should know the basics, but it's easy enough to learn that it doesn't really deserve a whole class, unless you're teaching students to build databases, which is definitely niche.
I'm sure it's needed by a large percentage of programmers, but maybe that's my bias speaking.
The relational algebra definitely fits into CS foundations, in terms of things like ACID and serializability etc.
That was a fun course where I learnt horribly bad patterns like "your data forms a key so you should use that as your key instead of having a surrogate key"
The prof was good at relational algebra, but not designing software
I totally understand looking for graduates that know SQL for certain positions, but it is a specialty not taught in any core class. You can absolutely know CS and entire areas of programming without it.
I used SQL all the time in web dev but haven't touched it since going to college. I do embedded so it's just never been necessary.
Now if this is a web dev or data related posting and students are coming in supposedly for this specialty then this ask HN is a little more understandable. Although perhaps poorly worded.
The way this was handled where I went to school the DB class was basically presented as a history of database technology class where the CS fundamentals behind how each type of database works, what it's weak points and strong points are, etc..
SQL was then something that was a side thing in the class that you were supposed to pick up for one of the projects.
We had to learn to normalize schemas in that class. I have yet to come across a team in my career where normalization was understood well enough to where it could be correctly matched to the architecture of the entire system.
The other amazing takeaway was going through the historical failures of hierarchical & graph databases. Since I started working hierarchical databases have come back with tons of hype twice now (once in the late 90s, once in the late 2010s). Both times they failed exactly the same way as history would predict based on what happened in the 60s and 70s that led to the development of relational databases. Sometimes it seems like the industry hype train is completely unaware of computer science.
The class I did was largely focused on relational algebra. So we learned the formal language of that (which I've largely forgotten) but did many practical exercises in SQL. I think it was valuable, since "thinking in sets" is something that people who only do linear algebra or algorithms might not be used to and so they carry a more linear problem-solving thought process with them.
I'm sure it's needed by a large percentage of programmers, but maybe that's my bias speaking.