expand.(): Cannot yet handle negatives

[arthur-albuquerque]

Hi, tidytable has changed the way I wrangle my data.

Unfortunately, I'm having issues using expand.() probably because my vector contains negative values.

"expand.(nesting(a3,b3,sig3),log_pred_mass = seq(-20,20,0.2))
Error in fsort(x, decreasing = decreasing, na.last = na.last) :
Cannot yet handle negatives."

Do you have any idea when it will handle negative values?

Thank you!

[markfairbanks]

Interesting - I didn't know that data.table::fsort() couldn't handle negatives.

Should be a relatively easy fix - I'll switch over to using base::sort(x, method = "radix"). It's a fast sort that data.table contributed to base R, but it also works on negative values.

[arthur-albuquerque]

That would be great. Thanks!

[markfairbanks]

reprex

pacman::p_load(tidytable)

test_df <- tidytable(x = c(1, 1, 2), y = c(1, 1, 2))

test_df %>%
  expand.(nesting.(x, y), z = c(-1.0, 1.0))
#> Error in fsort(x, decreasing = decreasing, na.last = na.last): Cannot yet handle negatives.

[markfairbanks]

Also I noticed in your example above you were using tidyr::nesting() instead of tidytable::nesting.(). Though it works, your code will run faster if you use the tidytable version:

pacman::p_load(tidytable, tidyverse)

data_size <- 1000000
test_df <- tidytable(
  x = sample(1:100, data_size, TRUE),
  y = sample(letters, data_size, TRUE)
)

bench::mark(
  tidytable_nesting = test_df %>% expand.(nesting.(x, y)),
  tidyr_nesting = test_df %>% expand.(nesting(x, y))
) %>%
  select.(expression, median)
#> # A tidytable: 2 × 2
#>   expression          median
#>   <bch:expr>        <bch:tm>
#> 1 tidytable_nesting   8.77ms
#> 2 tidyr_nesting      22.58ms

[arthur-albuquerque]

Oh wow, I didn't notice that! Great tip.

[markfairbanks]

Fixed! Thanks for catching this 👍

I'll submit the new version of tidytable to CRAN in the next few days so it will be fixed there as well.

library(tidytable)

test_df <- tidytable(x = c(1, 1, 2), y = c(1, 1, 2))

test_df %>%
  expand.(nesting.(x, y), z = c(-1.0, 1.0))
#> # A tidytable: 4 × 3
#>       x     y     z
#>   <dbl> <dbl> <dbl>
#> 1     1     1    -1
#> 2     1     1     1
#> 3     2     2    -1
#> 4     2     2     1