When to use pipes?

I’ve been thinking about pipes lately as @epogrebnyak and I decided to port the Causal Inference Mixtape (https://mixtape.scunning.com/) to Julia and the original R code is very pipe heavy. My first language was R but I use data.table and not pipes there and have never been a big fan of them. In julia I have used native pipes for one argument functions and think it is still easy to reason about the code. Currently I am looking at Chain.jl as a replacement for Magrittr pipes. Are there any opinions on this topic? Should we use pipes? Which package? How many functions make it worth it to use a pipe?
Thanks for your input!

There was a long thread about this recently How often do you use the |> operator?




I was searching for “pipe” but of course “|>” would have been good thanks!

It seems people use it for “hacking” in the REPL and some function composition. I have to think some more about how this translates to a general design decision for the Mixtape translation.

I now realize that this was suboptimal phrasing :rofl:

1 Like