Any JOIN is what the author is calling a "CROSS JOIN", with "ON" (which is optional) basically being a WHERE condition, and "LEFT", "RIGHT", "INNER", "OUTER" words only describing whether or not to drop a result if the left/right side of the join would be a NULL reference.
And, of course, a "UNION JOIN" is not a join at all, because nothing is joined.
SELECT a.pk
FROM table1 a, table2 b
WHERE a.pk *= b.fk
With LEFT JOIN:
SELECT posts.title,
GROUP_CONCAT(tags.tag) AS tagarr
FROM posts
LEFT JOIN post_tags ON post_tags.post_id = posts.id
LEFT JOIN tags ON tags.tag_id = post_tags.tag_id
SELECT posts.title,
GROUP_CONCAT(tags.tag) AS tagarr
FROM posts, post_tags, tags
WHERE posts.id *= post_tags.post_id
AND tags.tag_id *= post_tags.tag_id
ERROR 1064 (42000): You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near '= post_tags.post_id [...]The ANSI 89 syntax for inner joins is pretty ubiquitous, as it can be a little easier to write code generators that way.
Reason being that it's used to declare and restrict which table should be updated in the UPDATE section as opposed to optional joined tables listed in the FROM section.
PS: The *= is plateform specific however (have no clue which). Usually you just write a more SQL standard WHERE section.
https://www.postgresql.org/docs/13/tutorial-join.html
See also join_type: https://www.postgresql.org/docs/13/sql-select.html
Nearly all programmers get locked into thinking in terms of sequential, imperative programming, and the big differences are procedural vs OOP. They think that's always what programming looks like. They never learn to think in any other terms. They think functional programming is a weird outlier and looks like Lisp.
It's not entirely their fault. This is all they're typically taught in school. They're taught they don't need to think any other way than how Java, C, or C# tells them they should.
So they're not used to thinking in terms of sets. They're not used to declarative programming. Indeed, they'll often argue that anything other than general purpose, imperative programming isn't "real" programming. That's why Professor Brailsford from Computerphile got so much flack for saying that HTML was a programming language.
I'm not really interested in the argument, but all you've said so far is "it's domain specific" or "it's not Turing complete", neither of which are actually requirements for programming languages as a whole. It's not particularly difficult to devise a sufficiently loose definition of "program" that HTML fits into, too. Or, rather, it's difficult to envision a definition that LaTeX fits into and HTML+CSS fits into and SQL fits into and LISP fits into and so on, but HTML does not without your argument simply restating "it's domain specific" or "it's not Turing complete".
However, it's also not a particularly interesting discussion, either, because the only purpose behind such categorization is to assign value judgements like "a 'real' programming language can be used for systems programming".
And the other unstated purpose is gatekeeping. Oof, for such a relatively risk-free, comfortable and privileged profession, we're such an insecure lot.
And Regex isn't a programming language because it's not Turning complete?
It's natural to for me to say that writing a regular expression is programming a state machine.