NEWS.md

numberstrings (development version)

• Initial CRAN submission.
• This new package is a spin-off from scrutiny. Much of the code was copied from there; see lhdjung/scrutiny#65 on Github.
• Notably, however, scrutiny::split_by_parens() was not included here. It was replaced by separate_wider_parens(), which is faster and more consistent with tidyr's separate_wider_*() functions. Here is how it differs from split_by_parens():
  • The sep and check_sep arguments were renamed to delim and check_delim, consistent with tidyr::separate_wider_delim().
  • Columns to be split are no longer selected automatically. You need to specify them explicitly using the cols argument. If you want the same column selection as with split_by_parens()' default, use cols = everything(). Note that check_delim also influences column selection, just like check_sep in split_by_parens().
  • Edge cases are handled more gracefully. For example, any characters after the closing delimiter are stored in a new column. The name of this column is controlled by the new name_after argument. By contrast, split_by_parens() would simply ignore characters after the closing delimiter.
  • TODO: SAY HOW MUCH FASTER THE NEW FUNCTION IS COMPARED TO THE OLD ONE, BUT FIRST, ADD MISSING FUNCTIONALITY TO ENABLE A FAIR COMPARISON!
  • Some new arguments are passed down to separate_wider_regex().
  • By default, the new trim argument removes any spaces around the delimiters.
  • Internally, separate_wider_parens() uses tidyr::separate_wider_regex() as its workhorse instead of dplyr::mutate() with across().