pwwang

I would recommend [rtoml](https://github.com/samuelcolvin/rtoml) See: https://github.com/pwwang/toml-bench

> FYI Python is gaining `tomllib` in the 3.11 standard library, which is adapted from `tomli` `tomli` is also a good choice, not only because of the reason you mentioned,...

I like the second bullet point from the original proposal https://github.com/samuelcolvin/rtoml/issues/23#issuecomment-917640979. That is, to allow something like a `none_string` or `none_repr`, which defaults to `"null"`, to dump `None` into a...

Or UCSC genome browser?

I had the same problem. Not sure if this is your case, but in mine, I had too few cells in one of the groups marked by `labels` I passed...

`distinct()` is now optimized when it's on existing columns. Now the time spent for `a()`, `b()`, `ax()`, and `bx()`: ``` a(): 1.0249294309978723 b(): 0.7078308410054888 ax(): 6.315244514997175 bx(): 5.231293500000902 ```

What if people go back to some accomplished chapter? Or jump to future chapters (is that allowed)?

Like we need an LSP?

This feature is still experimental with dplyr. See: https://dplyr.tidyverse.org/reference/mutate.html https://dplyr.tidyverse.org/reference/summarise.html https://dplyr.tidyverse.org/reference/slice.html We will introduce it until it's stabilized.

Yes, that's what I wanted. ~~Just a naive thought about the recursion: how about using a sentinel variable?~~