Splitting strings into number and string (with missings)

Question

I am trying to separate numbers and characters in a column of strings. So far I have been using tidyr::separate for doing this, but am encountering errors for "unusual" cases.

Suppose I have the following data

df <- data.frame(c1 = c("5.5K", "2M", "3.1", "M"))

And I want to obtain a data frame with columns

data.frame(c2 = c("5.5", "2", "3.1", NA),
c3 = c("K", "M", NA, "M))

So far I have been using tidyr::separate

df %>%
separate(c1, into =c("c2", "c3"), sep = "(?<=[0-9])(?=[A-Za-z])")

But this only works for the first three cases. I realize this is because ?<=... and ?=... require the presence of the regex. How would one modify this code to capture the cases where the numbers are missing before the letters? Been trying to use the extract function too, but without success.

Edit: I suppose one solution is to break this up into

df$col2 <- as.numeric(str_extract(df$col1, "[0-9]+"))
df$col3 <- (str_extract(df$col1, "[aA-zZ]+"))

But I was curious whether were other ways to handle it.

Answer 1

extract(df, c1, into =c("c2", "c3"), "([\\.\\d]*)([a-zA-Z]*)")
#    c2 c3
# 1 5.5  K
# 2   2  M
# 3 3.1   
# 4      M

You can use seperate simply in this way, but there should be a more elegant method..

df %>% separate(c1, into =c("c2", "c3"), sep = "(?=[A-Za-z])")
#    c2   c3
# 1 5.5    K
# 2   2    M
# 3 3.1 <NA>
# 4        M