Re: Split string and escape delimiter

[Sean Conner <sean@...>]

It was thus said that the Great Minh Ngo once stated:
> There are a lot of examples on splitting a string on the net.
> What I'm looking for is the ability to include the delimiter in the strings.
> 
> The following desirable example with '\' as the escape character:
> 
> ````
> 'foo','bar','foo,bar' = split('foo,bar,foo\\,bar' , ',')
> ````

lpeg     = require "lpeg"
local Ct = lpeg.Ct
local C  = lpeg.C
local P  = lpeg.P

function split(s,delim)
  -- ----------------------------------------------------------------------
  -- Use the given delimeter, or if not supplied, use ':' as the delimeter.
  -- Change this if you want a different default.
  -- ----------------------------------------------------------------------

  local delim = delim or P":"

  -- ------------------------------------------------------
  -- Make sure our delimeter is an LPeg expression.
  -- ------------------------------------------------------

  if type(delim) == 'string' then
    delim = P(delim)
  end

  -- ---------------------------------------------------------------------
  -- A single character.  This can be '\' followed by any character (even
  -- the given delimeter) or a single character that doesn't match our
  -- delimeter.
  -- ---------------------------------------------------------------------

  local char  = P[[\]] * P(1)
              + (P(1) - delim)

  -- ----------------------------------------------------------------------
  -- A word is 0 or more characters.  After we have collected our word,
  -- remove any '\'.  There's a bug here (a corner case), but I'll leave
  -- that as an exercise for the reader to fix.
  -- ----------------------------------------------------------------------

  local word  = char^0 / function(s) s = s:gsub([[\]],'') return s end

  -- ----------------------------------------------------------------------
  -- a list of words is a word, followed by zero or more delimeter and word.
  -- ----------------------------------------------------------------------

  local words = word * (delim * word)^0

  -- --------------------------------------
  -- collect all the results into a table.
  -- --------------------------------------

  local match = Ct(words)
   
  return match:match(s)
end

print(unpack(split('foo,bar,foo\\,bar' , ',')))

foo     bar     foo,bar

  -spc (Is this a good enough example?)