Why R is better than Python? Maybe it is not and we should use both?

[Slavek]

Hi,
I have a general question.
Why do you think R is better than Python?

There are multiple comparisons on the internet but I would like to know your opinion as you are already R users, not people considering using this software for analyses or visualizations.

[pathos]

Benchmarks https://h2oai.github.io/db-benchmark (not the best examples, because dtplyr or tidytables exist, or Cython and Numba for Python, but good enough in my book)

Better implementation of C++ (Rcpp) and other languages, but my knowledge on this could be outdated and wrong.

Python can't run R, but R can run Python.

Library mention necessities (Python's pd.xyz vs. R's xyz or library::xyz)

And for me, the syntax being similar to human languages (that I know) is pretty nice -- subject first is common. For example:

df |>
  function(args)

subject |>
  verb(object/'adverbs' or the hows)

It's a crude example, but it works for me nicely.

Dependencies

Last but definitely not least, debug messages

[arthur.t]

• dplyr is more elegant and powerful than pandas. I find pandas really clunky. Even just indexing you have to use the .ix method to index. And sometimes a pivot or merge will produce hierarchical indexes, which is really confusing.
• the pipe operator in R is great and no equivalent in python
• python is a messy mix of functional programming and object oriented programming while R is more purely functional. I find it difficult to remember in python if I need to sort(object) or object.sort()
• python has way too many variable types. All those types leads to lots of debugging and head scratching. Just one example, map in R produces a list. map in python returns an "iterator object". I just want a list!
• I like RStudio as an IDE more than PyCharm, VSCode, or Spyder
• R notebooks much more customizable and presentable than Jupyter. I can't understand how it became acceptable to share messy Jupyter html outputs with code and output with non-programmer clients.
• Python debuggers still use ipython which is so clunky and archaic. And because python has such a mess of data types and packages, I spend a lot of time in the debugger...
• Much more straightforward in R to install software and packages. I hate when I have to mess around with pip and conda.

[andresrcs]

Both have their strengths and weaknesses, so the best approach is to choose the right tool for the job at hand at any given moment, even including tools other than R and Python.