From: "Ara.T.Howard" <Ara.T.Howard@...>
Date: 2005-10-30T09:06:58+09:00
Subject: Re: csv.rb a start on refactoring.

On Sun, 30 Oct 2005, James Edward Gray II wrote:

> On Oct 29, 2005, at 12:11 PM, Ara.T.Howard wrote:
>
>> ==========================================
>> CSV2[8] => FAILED (RuntimeError)
>> ==========================================
>> input:
>> "a,,,"
>> csv:
>> ["a", "", "", ""]
>> expected:
>> ["a", nil, nil, nil]
>> ==========================================
>
> Which one of those is more correct must be debatable.  I didn't see a concept 
> of nil in the RCF document and nil stringifies to "", of course.  Regardless, 
> it's a simple fix:
>
> module CSV2
>  def self::parse_line( line )
>    csv = Array.new
>    line.scan(/\G(?:^|,)(?:"((?>[^"]*)(?>""[^"]*)*)"|([^",]*))/) do
>      csv << if $1.nil?
>        if $2 == ""
>          nil
>        else
>          $2
>        end
>      else
>        $1.gsub('""', '"')
>      end
>    end
>    csv
>  end
> end
>
> That gives us two less errors.

i think you need both - otherwise

   a,,,

and

   a,"","",""

cannot be told apart - but they are clearly different...  debateable for sure.

the rfc is on your side here it seems:

   field = (escaped / non-escaped)

   escaped = DQUOTE *(TEXTDATA / COMMA / CR / LF / 2DQUOTE) DQUOTE

   non-escaped = *TEXTDATA

so '""' and '' are the same in the rfc.  seems an over-sight though since it's
impossible to mark 'empty' differently than 'missing'.  in any case it'd
probably be good to maintain backwards compat here so as not to break any apps
- in that case nil would win.


>> ==========================================
>> CSV2[9] => FAILED (RuntimeError)
>> ==========================================
>> input:
>> ","
>> csv:
>> [""]
>> expected:
>> [nil, nil]
>> ==========================================
>
> You're input there was strip()ed.  It was actually " , ".  The expression 
> obviously fails this case either way, but in the RCF I read:
>
> "4...  Spaces are considered part of a field and should not be ignored."

i stripped input simply so i could write

   %w( , )

instead of

   %w(,)

which is a little cryptic.  i did mean for it to be ',' - no spaces.  sorry if
that was confusing.  note that each and every input is stripped by design -
but only for clarity.

> I would interpret that to mean that the correct solution is [" ", " "].  Can
> you please explain why that's wrong?

taking into account the stripping above a correct solution should be

   [nil, nil]

or possibly

   ["", ""]

but it should surely be two fields either way.  i guess we are saying the same
thing here...


so, except for blowing up on trailing commas the re seems to be pretty dang
close.  impressive.

cheers.

-a
-- 
===============================================================================
| email :: ara [dot] t [dot] howard [at] noaa [dot] gov
| phone :: 303.497.6469
| anything that contradicts experience and logic should be abandoned.
| -- h.h. the 14th dalai lama
===============================================================================