TysonStanley/tidyfast

Outadate readme regarding dt_separate?

eliocamp opened this issue · 1 comments

The readme currently states that

Testing with a 4 MB data set with one variable that has columns of “A.B” repeatedly, shows that dt_separate() is fast but less memory efficient than tidyr::separate().

But the benchmark shows:

#> # A tibble: 3 × 3
#>   expression            median mem_alloc
#>   <chr>               <bch:tm> <bch:byt>
#> 1 separate               4.98s    3.89GB
#> 2 dt_separate             73ms   22.92MB
#> 3 dt_separate-mutable  65.91ms   19.09MB

Which, if I'm reading correctly, makes dt_separate orders of magnitude more memory efficient than tidyr::separate!

Good catch! That was definitely part of an update that went unnoticed in the README. Thanks for the issue.