*_join syntax inconsistency?

[rdboyes]

When you do a join in TidierData on two columns that do not have the same name, the syntax we use is e.g.:

@left_join(d1, d2, "c2" = "c1")

Does anyone remember why we chose this syntax? It's not the same as DataFrames leftjoin (leftjoin(d1, d2, on = :c1 => :c2) and it's also not the same as joins in R (left_join(d1, d2, by = c("c1" = "c2")) or join_by(c1 == c2)).

I fully recognize that this may be my own fault since I remember doing some of the initial work on the joins, but I think we should fix this to be consistent with R, probably going with the join_by version since that is the most up-to-date syntax.

[drizk1]

I was still learning tidyverse when joins were implemented so im not sure.

that being said, with ~ 10 more lines in the parse_join_by function, i was able to add support for the following without breaking any code and im sure theres a better way to do it
ex df from df.jl docs

@inner_join(a, b, join_by(City==Location, Job==Work))
@inner_join(a, b, join_by("City"=="Location", "Job"=="Work"))
@inner_join(a, b, join_by(:City=>:Location))