Matters what you're using the language for more than which language is being used I would imagine.
Scala users make more than .NET users, but that isn't because Scala is more in-demand, it's because that's the language of choice for a lot of Machine Learning / Big Data folks using spark who are highly sought after. I would bet that average Scala and Java compensation is very different, with Scala being much higher, due to the Java average compensations getting watered down by salaries of Java developers that just make CRUD applications but at the top of the scale they're likely similar and they both map back to Spark developers that are Machine Learning / Data Scientist / Big Data engineers.
Same with SQL- in general it isn't highly compensated because "everyone" knows/uses SQL (rarely well, though)- however there are also tons of Data Engineers whose core competency is SQL and they make great money.
The median pay might be quite low for Java, because there are also a hojillion poorly-paid enterprise wageslaves toiling away in the cubicle farms.
But the absolute number of people earning a lot (100k, 200k, whatever you think that is) might well be higher for Java than for anything else, because it's used in some significant big tech companies and in finance.
So, if you're of average ability, learning Java might not juice your pay much, but if you're highly talented, it might.
I don’t think that’s well specified enough to measure. But assuming you mean something like median comp of a job that requires X, I’d imagine the top of the list would all be things that are pretty obscure.
Suppose there’s one hedge fund that wants an expert in k and is willing to pay whatever it takes. Does that mean people should go out and learn k?
