I work in data science, and last year learned the Scala/Spark stack, switching from Python's numeric computing/ML ecosystem. I would say that this involved some

I work in data science, and last year learned the Scala/Spark stack, switching from Python's numeric computing/ML ecosystem. I would say that this involved something like 200 hours of work to get to the point where environment differences were no longer a daily distraction, and of that, only about 20 hours were due to the language itself–most of it was things like getting an IDE to work, futzing around with sbt, learning about the common libraries, Spark configuration headaches etc. 200 hours is a lot of time. It was worth it in this case, because Spark lets you do a lot of things that the Python environment doesn't (and vice versa), but I would be pretty hesitant to invest that time for something without an obvious benefit.