On Sun, May 6, 2012 at 8:00 PM, Donald Griggs <dfgri...@gmail.com> wrote: > Regarding: What precisely are the > "improvements" in handling of CSV inputs? > > > Gabor, I don't know about "precisely" -- I'll let others on the list tell > me where I'm off, but here's my take: > > > A lot of strange things call themselves csv, but the change attempts to > make the sqlite3 utility's CSV inputs perform a bit more closely to RFC4180. > http://tools.ietf.org/html/rfc4180 > > http://en.wikipedia.org/wiki/Comma-separated_values#Toward_standardization > > In particular, during CSV mode import: > -- Allow any field to be surrounded by double quote characters without > those characters being considered part of the field data. > -- Allow fields to contain embedded commas (or other separators) when the > field is surrounded by double quote characters. > -- Allow fields to span multiple lines if they are surrounded by double > quote characters. > -- Allow the double quote character to be escaped by having two adjacent > double quote characters. (But note that a field consisting solely of two > double quote characters still represents an empty string field.) > > -- On output in CSV mode, surround text fields with double quotes when > needed. > > > See check-in [93aa17d866] http://www.sqlite.org/src/info/93aa17d866 >
Thanks very much. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users