When I got annoyed in my own work that some data wrangling syntax was repetitive I was just really glad that I could easily build my optimal solution and didn't have to just accept that there's one suboptimal (for me) way. In Python and R, if you like what they offer that's good, if not - not so good.
Part of the problem comes from Julia not being geared towards DataFrames like R is, but I gladly trade a bit of convenience in one domain against a lot of expressive freedom with very clean rules that apply everywhere.
For example, I think it's quite good that you can only have "weird" behavior in Julia with macros, but they give you a visual indicator with the @ that you're seeing non-standard syntax. While in R, the non-standard evaluation means that literally anything could happen to the variables you pass into any function. It makes for some convenient syntax in some cases, yes, but it's so confusing as a system for writing software! You never really know if you're looking at a variable or just a name, for example.
One thing that I’ve recently seen which concerns me long term is the creation of various competing macro syntaxes for reducing the wordiness of Julia. There are many competing implementations of pipes and other syntax sugars. These macros definitely make things easier, but as you use them the code becomes more difficult for another to understand and since there is at this time, no one set of macros to use, you’ll have to know each of the competing sets to make since of examples.