Discussion about this post

User's avatar
Daniel Gimenez's avatar

The `collapse` package uses pipe tidyverse-like functions that was written in C, like data.table. So, in R you can code with same intuitive syntax than dplyr but with code performance as good as (or even better than) data.table.

Collapse and other very fast packages are gathered in a meta-package called "fastverse".

According to current benckmarks, collapse is faster than Polars in most of data manipulation operations: https://github.com/AdrianAntico/Benchmarks

Expand full comment
Shibaprasad Bhattacharya's avatar

"This is to say that it has gone mainstream. And when something goes mainstream, people don’t necessarily use it because it’s the best tool for the job; more often than not they use it because everyone else is using it."

This is a big reason why I use Python, along with the fact that productionizing things is comparatively easier with Python. For any work where I don’t expect my teammates to collaborate with me, I use R.

Expand full comment

No posts