R, split data frame column into 2 on last occurrence of either a + or a – sign

We are attempting to split 2 dataframe columns using the last occurrence of either a + or – sign, splitting the string immediately before the sign.

What we have

zed = data.frame(
  value = c('+23-25', '-23.5+25.5', '-24+24', '-3-6'),
  label = c('p34-150', 'p 243-130', 'p215 +115', 'p215 +115'),
  stringsAsFactors = FALSE
)

What we are going for

output = data.frame(
  value1 = c('+23', '-23.5', '-24', '-3'),
  value2 = c('-25', '+25.5', '+24', '-6'),
  label1 = c('p34', 'p 243', 'p215', 'p215'),
  label2 = c('-150', '-130', '+115', '+115'),
  stringsAsFactors = FALSE
)

separate(value, into = c("value1", "value2"), sep = "+|-") doesn’t quite work, +|- is an invalid regular expression. "(?=[+-][^+-]*$)[+-]" not working as intended either, returns empty columns. At the surface, it seemed like this would be simple but we are struggling.

Edit: perhaps this requires a separate posting, but we have a weird issue with an irregular space, causing the good solution below to not work:

> zed$Total[1]
[1] "O 239.5−110"
> head(zed %>% separate(Total, into = c("label1", "label2"), sep = "(?=[\\+-][^\\+-]*$)"), 1)
        label1  label2
1  O 239.5−110    <NA>
Warning message:
Expected 2 pieces. Missing pieces filled with `NA` in 12 rows [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]. 
> gsub(' ', '', zed$Total[1])
[1] "O 239.5−110"

Cannot gsub away the space after the O, for some reason this causes the separate() to not work.

>Solution :

You were close with your last attempt; you just need to escape the + and don’t need the [+-] at the end.

library(tidyr)

zed %>%
  separate(
    value, 
    into = c("value1", "value2"), 
    sep = "(?=[\\+-][^\\+-]*$)"
  ) %>%
  separate(
    label, 
    into = c("label1", "label2"), 
    sep = "(?=[\\+-][^\\+-]*$)"
  ) 

  value1 value2 label1 label2
1    +23    -25    p34   -150
2  -23.5  +25.5  p 243   -130
3    -24    +24  p215    +115
4     -3     -6  p215    +115