On Jan 6, 2009, at 6:14 PM, sqlite-users-requ...@sqlite.org wrote: > delete from talks where exists > (select 1 from talks t2 > where talks.member_id = t2.member_id and talks.date = t2.date and > talks.rowid > t2.rowid);
Igor, this worked fabulously, thank you very much. I also tried your other routine: > delete from talks where rowid in > (select t2.rowid from talks t2 > where talks.member_id = t2.member_id and talks.date = t2.date > order by t2.rowid offset 1); But I could not get anything to function, even when isolating the second SELECT, I continued to get error messages. I am working in the Mac OS X environment, with SQLite version 3.4.0, if that makes any difference. Alexey, thank you very much for your idea to put a CONSTRAINT on the table in the first place, that is the trick for a long term solution. Here is how I have put it together: CREATE TABLE talks (member_id INTEGER, date DATE, CONSTRAINT constraint_ignore_dup UNIQUE (member_id, date) ON CONFLICT IGNORE); I believe that I understand this statement, except for the term constraint_ignore_dup. Is that a variable name? Could it be pretty much anything I want, and if so, what is its purpose? Thank you to all the responders; you are the nicest internet contacts I have ever made. Craig Smith cr...@macscripter.net _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
