pivot longer error message : can't combine double and characters

[zemickeyman]

Hello,

I've got the following table (example)

I would like to pivot it longer using the columns prefixes A and B transfered in a new column named treatment.
But when I try to pivot it longer , I've got the followng error message :
Error in pivot_longer():
! Can't combine A_length and A_chaine

Here is my code:
to create the wide table:

tableau_2 <- tibble(
  lot = c(25789, 24856),
  temp_ID = c("ppp","ooooo"),
  A_length = c(45, 25),
  A_volume = c(25, 856),
  A_weight = c(236, 4568),
  A_chaine = c("bla", "bli"),
  B_length = c(63, 47),
  B_volume = c(569, 145),
  B_weight = c(12, 28), 
  B_chaine = c("blo", "blu")
)

then to try pivoting it

tableau_1 <- tableau_2 %>%
  pivot_longer(
    cols = -c(lot,temp_ID),
    names_to = c("treatment", "measure"),
    names_sep = "_",
    values_to = "value"
  ) 

I don't really understand what is the problem
could you help me please ?

Mickael

[margusl]

For this you could use ".value" in names_to:

library(tidyr)
library(dplyr)

tableau_2 %>%
  pivot_longer(
    cols = -c(lot,temp_ID),
    names_to = c("treatment", ".value"),
    names_sep = "_",
  )
#> # A tibble: 4 × 7
#>     lot temp_ID treatment length volume weight chaine
#>   <dbl> <chr>   <chr>      <dbl>  <dbl>  <dbl> <chr> 
#> 1 25789 ppp     A             45     25    236 bla   
#> 2 25789 ppp     B             63    569     12 blo   
#> 3 24856 ooooo   A             25    856   4568 bli   
#> 4 24856 ooooo   B             47    145     28 blu

In your current attempt you try to fill treatment & measure columns from existing column names and combine numeric and character objects in 3rd column, value . This would work if you'd convert numerics to characters first, but likely it is not what you are after:

tableau_2 %>%
  mutate(across(matches("^(A|B)_"), as.character)) %>%
  pivot_longer(
    cols = -c(lot,temp_ID),
    names_to = c("treatment", "measure"),
    names_sep = "_",
    values_to = "value"
  ) 
#> # A tibble: 16 × 5
#>      lot temp_ID treatment measure value
#>    <dbl> <chr>   <chr>     <chr>   <chr>
#>  1 25789 ppp     A         length  45   
#>  2 25789 ppp     A         volume  25   
#>  3 25789 ppp     A         weight  236  
#>  4 25789 ppp     A         chaine  bla  
#>  5 25789 ppp     B         length  63   
#>  6 25789 ppp     B         volume  569  
#>  7 25789 ppp     B         weight  12   
#>  8 25789 ppp     B         chaine  blo  
#>  9 24856 ooooo   A         length  25   
#> 10 24856 ooooo   A         volume  856  
#> 11 24856 ooooo   A         weight  4568 
#> 12 24856 ooooo   A         chaine  bli  
#> 13 24856 ooooo   B         length  47   
#> 14 24856 ooooo   B         volume  145  
#> 15 24856 ooooo   B         weight  28   
#> 16 24856 ooooo   B         chaine  blu

[zemickeyman]

Hello Margusl,

many thanks for your help.

Micka