I also work in the data space.<p>Yes.<p>Stepping back from the Rust specific things for a moment. I’ll go a step further and say “you should take the opportunity to learn multiple, different languages”.
At worst you learn a new language and grow your skills.
Different languages have different strengths/weaknesses and styles, and being exposed to them gives you more “tools” to use to solve problems, and different styles give you different ways to “think” which is useful for solving tricky problems (Haskell and Lisp are good examples of this) that may be tackled in much simpler ways of you swap to different styles/languages.<p>Now, as for Rust specifically: it’s a great language, it’s popular, it’s performant and has a number of other benefits. It’s fast enough, that you can write reasonably normal, idiomatic code in it and get shockingly good performance; why do I bring this up? Without too much effort, you can write code that is correct (in ways much stronger than what python can offer) while giving you performance close or better than those libraries. You’re no longer restricted by what’s prepackaged in libraries - if you need something that isn’t available, building it, and getting good flexibility and performance.<p>Personally, I think Rust has a lot of potential for the data space. Previously you had to rely on expertly written and tightly optimised Java/Scala or C/C++, but Rust is opening up so much optionality.
Polars has basically jumped straight to the top of dataframe library performance, DataFusion/Ballista are shaping up to be viable replacement for Spark with better performance and less general pain. Libs like SQLX and Cornucopia are offering compile time type checked SQL in awesome ways. The data space is still somewhat young in Rust, but it has a huge amount of potential IMO.