datawizard versions of separate() and unite()

[jmgirard]

I'm hoping to retool my packages soon to use datawizard instead of dplyr and tidyr. However, a few things I need that don't seem to exist yet are datawizard versions of separate() and unite() from tidyr. Any plans to create these?

[etiennebacher]

I think unite() should be fairly easy to implement, but separate() is a bit trickier because of the arguments fill and extra. I started working on implementing separate() in poorman (cf nathaneastwood/poorman#107) but it's still WIP (and I don't know when I'll get back at it). It might provide a starting point if someone else wants to do it.

[strengejacke]

library(datawizard)

d <- data.frame(
  x = c(NA, 1:3),
  y = c(letters[1:3], NA_character_),
  z = 6:9,
  m = c("X", NA_character_, "Y", "Z")
)
d
#>    x    y z    m
#> 1 NA    a 6    X
#> 2  1    b 7 <NA>
#> 3  2    c 8    Y
#> 4  3 <NA> 9    Z

data_unite(d, new_column = "xyz")
#>         xyz
#> 1  NA_a_6_X
#> 2  1_b_7_NA
#> 3   2_c_8_Y
#> 4  3_NA_9_Z
data_unite(d, new_column = "xyz", remove_na = TRUE)
#>        xyz
#> 1    a_6_X
#> 2    1_b_7
#> 3  2_c_8_Y
#> 4    3_9_Z

data_unite(d, new_column = "x")
#>           x
#> 1  NA_a_6_X
#> 2  1_b_7_NA
#> 3   2_c_8_Y
#> 4  3_NA_9_Z
data_unite(d, new_column = "x", remove_na = TRUE)
#>          x
#> 1    a_6_X
#> 2    1_b_7
#> 3  2_c_8_Y
#> 4    3_9_Z

data_unite(d, new_column = "x", append = TRUE)
#> The name for `new_column` already exists as variable name in the data.
#>   This variable will be replaced by `new_column`.
#>           x    y z    m
#> 1  NA_a_6_X    a 6    X
#> 2  1_b_7_NA    b 7 <NA>
#> 3   2_c_8_Y    c 8    Y
#> 4  3_NA_9_Z <NA> 9    Z
data_unite(d, new_column = "x", remove_na = TRUE, append = TRUE)
#> The name for `new_column` already exists as variable name in the data.
#>   This variable will be replaced by `new_column`.
#>          x    y z    m
#> 1    a_6_X    a 6    X
#> 2    1_b_7    b 7 <NA>
#> 3  2_c_8_Y    c 8    Y
#> 4    3_9_Z <NA> 9    Z

data_unite(d, new_column = "xyz", append = TRUE)
#>    x    y z    m       xyz
#> 1 NA    a 6    X  NA_a_6_X
#> 2  1    b 7 <NA>  1_b_7_NA
#> 3  2    c 8    Y   2_c_8_Y
#> 4  3 <NA> 9    Z  3_NA_9_Z
data_unite(d, new_column = "xyz", remove_na = TRUE, append = TRUE)
#>    x    y z    m      xyz
#> 1 NA    a 6    X    a_6_X
#> 2  1    b 7 <NA>    1_b_7
#> 3  2    c 8    Y  2_c_8_Y
#> 4  3 <NA> 9    Z    3_9_Z

data_unite(d, new_column = "x2")
#>          x2
#> 1  NA_a_6_X
#> 2  1_b_7_NA
#> 3   2_c_8_Y
#> 4  3_NA_9_Z
data_unite(d, select = c("x", "z"), new_column = "new")
#>      y    m  new
#> 1    a    X NA_6
#> 2    b <NA>  1_7
#> 3    c    Y  2_8
#> 4 <NA>    Z  3_9
data_unite(d, select = c("x", "z"), new_column = "new", append = TRUE)
#>    x    y z    m  new
#> 1 NA    a 6    X NA_6
#> 2  1    b 7 <NA>  1_7
#> 3  2    c 8    Y  2_8
#> 4  3 <NA> 9    Z  3_9
data_unite(d, select = c("x", "z"), new_column = "new", append = TRUE, remove_na = TRUE)
#>    x    y z    m new
#> 1 NA    a 6    X   6
#> 2  1    b 7 <NA> 1_7
#> 3  2    c 8    Y 2_8
#> 4  3 <NA> 9    Z 3_9

data_unite(d, new_column = "xyz", separator = ".")
#>         xyz
#> 1  NA.a.6.X
#> 2  1.b.7.NA
#> 3   2.c.8.Y
#> 4  3.NA.9.Z

^{Created on 2023-05-25 with reprex v2.0.2}

[strengejacke]

@jmgirard data_unite() is implemented, feel free to test. data_separate() will follow in a separate (haha) PR.