I once also had this misconception that the CSV spec (which, as far as I can tell, exists only in the form of folklore) used quotes to distinguish character from numeric. But, if you've ever imported CUSIPs into Excel, you've seen this fallacy for what it is. Quotes are "properly" applied only to items with embedded separator characters (and maybe embedded newlines - see "http://supercsv.sourceforge.net/csvSpecification.html";).
Again, since there is no definitive spec (see " http://en.wikipedia.org/wiki/Comma-separated_values), this is subject to debate and varying interpretations but I'm pretty sure what I've outlined above is the minimal requirement. I find the promiscuous quoting of "writecsv" slightly annoying but it's acceptable. On Fri, Jan 8, 2010 at 10:49 PM, McGuinness, Brian <mcgui...@fdu.edu> wrote: > Comments on the J distribution j602/2008-03-03/16:45: > ... > 2. The writecsv verb in j603/system/packages/files/csv.ijs > puts quotes around > every data item, even numbers, when only character data > should be quoted. The > following code, which could probably be improved, seems to > quote the data > properly: > ... > --- Brian > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > -- Devon McCormick, CFA ^me^ at acm. org is my preferred e-mail ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
