library(tidyverse)
# -------------------------------------------------------------------------
# here I create a long table where I filter out negative occurences
spe_long <- tibble(
  ID = rep(1:100, each = 20),
  TAXON = paste0('spe', rep(1:100, 20)),
  `P/A` = sample(c(0,1), 2000, prob= c(7,3), replace = T)
) %>% filter(`P/A` != 0) 

spe_long # here it is 

# -------------------------------------------------------------------------
# now I want to convert this long table into wide formate
# the function needs two parameters, i.e. where to take variables names_from and where to take ~cover~ values_from
spe_long %>% pivot_wider(names_from = TAXON, 
                         values_from =`P/A`) 
# be aware that everything else is just omitted from spreading into columns, which in our case applies for IDs
# this output produces NAs; to get no nas, we have to add one more argument
wide <- spe_long %>% pivot_wider(names_from = TAXON, 
                         values_from =`P/A`,
                         values_fill = 0) # I add just 0 here, because my columns are of type 'double', if they were character, I would have to put there '0' or "0"

wide # here we have the output
# the IDs are however still present here 
# we can remove them using select function
wide %>% dplyr::select(-ID)

# -------------------------------------------------------------------------
# and if you want to get back to the long form, here is the function
# wide still has IDs here, becase we didnt store the line 28 into a file
wide %>% pivot_longer(-'ID') # this says: "put all values into one column and names into other column, but skip 'ID'
# alternatively
spe_long_again <- wide %>% pivot_longer(-1)

# -------------------------------------------------------------------------
# however! the table is now almost 10000 rows long, which is way more than spe_long that we used on the beginning
spe_long_again
spe_long
# it is because we have the 0 there, we can filter them out like this:
spe_long_again %>% filter(value != 0)