You may find something useful here

Get all of the quoted elements first and then get the comma-sep elements then put them back together.

I wrote a CSV-to-list converter a few years ago. I don't know if it's what you want.

Applescript:

(* Assumes that the CSV text adheres to the convention:
   Records are delimited by LFs or CRLFs (but CRs are also allowed here).
   The last record in the text may or may not be followed by an LF or CRLF (or CR).
   Fields in the same record are separated by commas (unless specified differently by parameter).
   The last field in a record must not be followed by a comma.
   Trailing or leading spaces in unquoted fields are not ignored (unless so specified by parameter).
   Fields containing quoted text are quoted in their entirety, any space outside them being ignored.
   Fields enclosed in double-quotes are to be taken verbatim, except for any included double-quote pairs, which are to be translated as double-quote characters.
       
   No other variations are currently supported. *)

on csvToList(csvText, implementation)
   -- The 'implementation' parameter must be a record. Leave it empty ({}) for the default assumptions: ie. comma separator, leading and trailing spaces in unquoted fields not to be trimmed. Otherwise it can have a 'separator' property with a text value (eg. {separator:tab}) and/or a 'trimming' property with a boolean value ({trimming:true}).
   set {separator:separator, trimming:trimming} to (implementation & {separator:",", trimming:false})
   
   script o -- Lists for fast access.
       property qdti : getTextItems(csvText, "\"")
       property currentRecord : {}
       property possibleFields : missing value
       property recordList : {}
   end script
   
   -- o's qdti is a list of the CSV's text items, as delimited by double-quotes.
   -- Assuming the convention mentioned above, the number of items is always odd.
   -- Even-numbered items (if any) are quoted field values and don't need parsing.
   -- Odd-numbered items are everything else. Empty strings in odd-numbered slots
   -- (except at the beginning and end) indicate escaped quotes in quoted fields.
   
   set astid to AppleScript's text item delimiters
   set qdtiCount to (count o's qdti)
   set quoteInProgress to false
   considering case
       repeat with i from 1 to qdtiCount by 2 -- Parse odd-numbered items only.
           set thisBit to item i of o's qdti
           if ((count thisBit) > 0) or (i is qdtiCount) then
               -- This is either a non-empty string or the last item in the list, so it doesn't
               -- represent a quoted quote. Check if we've just been dealing with any.
               if (quoteInProgress) then
                   -- All the parts of a quoted field containing quoted quotes have now been
                   -- passed over. Coerce them together using a quote delimiter.
                   set AppleScript's text item delimiters to "\""
                   set thisField to (items a thru (i - 1) of o's qdti) as string
                   -- Replace the reconstituted quoted quotes with literal quotes.
                   set AppleScript's text item delimiters to "\"\""
                   set thisField to thisField's text items
                   set AppleScript's text item delimiters to "\""
                   -- Store the field in the "current record" list and cancel the "quote in progress" flag.
                   set end of o's currentRecord to thisField as string
                   set quoteInProgress to false
               else if (i > 1) then
                   -- The preceding, even-numbered item is a complete quoted field. Store it.
                   set end of o's currentRecord to item (i - 1) of o's qdti
               end if
               
               -- Now parse this item's field-separator-delimited text items, which are either non-quoted fields or stumps from the removal of quoted fields. Any that contain line breaks must be further split to end one record and start another. These could include multiple single-field records without field separators.
               set o's possibleFields to getTextItems(thisBit, separator)
               set possibleFieldCount to (count o's possibleFields)
               repeat with j from 1 to possibleFieldCount
                   set thisField to item j of o's possibleFields
                   if ((count thisField each paragraph) > 1) then
                       -- This "field" contains one or more line endings. Split it at those points.
                       set theseFields to thisField's paragraphs
                       -- With each of these end-of-record fields except the last, complete the field list for the current record and initialise another. Omit the first "field" if it's just the stub from a preceding quoted field.
                       repeat with k from 1 to (count theseFields) - 1
                           set thisField to item k of theseFields
                           if ((k > 1) or (j > 1) or (i is 1) or ((count trim(thisField, true)) > 0)) then set end of o's currentRecord to trim(thisField, trimming)
                           set end of o's recordList to o's currentRecord
                           set o's currentRecord to {}
                       end repeat
                       -- With the last end-of-record "field", just complete the current field list if the field's not the stub from a following quoted field.
                       set thisField to end of theseFields
                       if ((j < possibleFieldCount) or ((count thisField) > 0)) then set end of o's currentRecord to trim(thisField, trimming)
                   else
                       -- This is a "field" not containing a line break. Insert it into the current field list if it's not just a stub from a preceding or following quoted field.
                       if (((j > 1) and ((j < possibleFieldCount) or (i is qdtiCount))) or ((j is 1) and (i is 1)) or ((count trim(thisField, true)) > 0)) then set end of o's currentRecord to trim(thisField, trimming)
                   end if
               end repeat
               
               -- Otherwise, this item IS an empty text representing a quoted quote.
           else if (quoteInProgress) then
               -- It's another quote in a field already identified as having one. Do nothing for now.
           else if (i > 1) then
               -- It's the first quoted quote in a quoted field. Note the index of the
               -- preceding even-numbered item (the first part of the field) and flag "quote in
               -- progress" so that the repeat idles past the remaining part(s) of the field.
               set a to i - 1
               set quoteInProgress to true
           end if
       end repeat
   end considering
   
   -- At the end of the repeat, store any remaining "current record".
   if (o's currentRecord is not {}) then set end of o's recordList to o's currentRecord
   set AppleScript's text item delimiters to astid
   
   return o's recordList
end csvToList

-- Get the possibly more than 4000 text items from a text.
on getTextItems(txt, delim)
   set astid to AppleScript's text item delimiters
   set AppleScript's text item delimiters to delim
   set tiCount to (count txt's text items)
   set textItems to {}
   repeat with i from 1 to tiCount by 4000
       set j to i + 3999
       if (j > tiCount) then set j to tiCount
       set textItems to textItems & text items i thru j of txt
   end repeat
   set AppleScript's text item delimiters to astid
   
   return textItems
end getTextItems

-- Trim any leading or trailing spaces from a string.
on trim(txt, trimming)
   if (trimming) then
       repeat with i from 1 to (count txt) - 1
           if (txt begins with space) then
               set txt to text 2 thru -1 of txt
           else
               exit repeat
           end if
       end repeat
       repeat with i from 1 to (count txt) - 1
           if (txt ends with space) then
               set txt to text 1 thru -2 of txt
           else
               exit repeat
           end if
       end repeat
       if (txt is space) then set txt to ""
   end if
   
   return txt
end trim

set csvText to "caiv2,2010BBDGRC,\"President, Board of Directors\"" & linefeed & "Another line, for demo purposes"
csvToList(csvText, {})
--> {{"caiv2", "2010BBDGRC", "President, Board of Directors"}, {"Another line", "for demo purposes"}}

Edit: It's been pointed out to me that the original script omitted empty fields. That's now corrected and the trim() hander's a trifle faster than it was. The script also now conforms to RFC 4180, but has options to trim leading and trailing spaces from unquoted fields — csvToList(csvText, {trimming:true}) — and/or to recognise a different field separator — eg. csvToList(csvText, {separator:tab, trimming:true}). An empty record gets the defaults: unquoted fields not trimmed; comma separator. Returns are now tolerated as record separators.

Last edited by Nigel Garvey (2012-03-29 11:51:52 am)