On Thu, Mar 8, 2012 at 11:22 AM, Nico Williams <n...@cryptonector.com> wrote: > I use this all the time as it's the only way in SQLite3 to do UPDATEs > with JOINs. I use it with an INTEGER PRIMARY KEY column. SQLite3 > tables always have a rowid somehow that's suitable for this, but it's > not a good assumption to bake into an application -- the application > should make sure that there's an INTEGER PRIMARY KEY or other suitable > PRIMARY KEY or UNIQUE columns.
I should clarify that this idiom works for: a) using a JOIN as a source of some of the values to update, b), if the table/view has an INSTEAD OF INSERT trigger then that can turn the insert of a row in a single table/view into an insert/update into multiple tables/views. I find this a very useful pattern. Nico -- _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
