Thanks for the suggestion about set theory.

Thanks for the follow-up comment, and I agree that a different visualization would be helpful here. I’m now thinking about a way where we could use layers in order to better picture the join syntax.

I know how actual tables are represented in computers. The reason I
pointed is when I was looking closely at those Ven diagrams for the
first col second fig or last col 2 fig the last line states where A.key
= null . But I could not figure that from the ven diagram. I switched
to a “tabular” diagram and it popped. Thanks for taking the effort to
explain it in the first place. sorry but I did not mean to be
commenting negatively.

But yeah, math class is hard.

Tables are not square – they are actually data. One common visualization of them uses boxes with square corners, others use boxes with rounded corners, and others use circles. “Circles and lines”, the lingo of some trolls originates from these earlier data models of tables and relationships.

LMGTFY https://www.google.ca/search?q=set+theory

Well, “actual tables” are a logical representation of data that resides on a physical disk, and the concept of a table itself is meant only for humans to comprehend. A computer doesn’t say “this is a table, and it’s a square”. It’s just bits on a page on a disk.

I guess I’m used to standard Venn diagrams where circles represent sets. And a circle works just as well as anything. Feel free to redraw the diagrams using squares, I’d be interested in seeing lots of different shapes used. After many years of mathematics I’ve never thought about using anything but circles to represent sets.

Thanks for the comment.