On Nov 16, 2013, at 11:02 PM, Kees Nuyt <k.n...@zonnet.nl> wrote: > For the application, the merge would look like a single > INSERT INTO merge_t statement.
Hmmmm…. clever lateral thinking, but I doubt this will fly in practice :) Two main issues: (1) ‘or ignore’ is most likely inappropriate as unrelated constraint violations will trigger it (e.g. null constraint). Resulting in a no-op altogether as neither insert nor update will do anything in practice. Leading to loss of data. (2) Both insert & update statement will execute irrespectively of the state of the data, doubling the workload. And, really, at this point (additional views, instead of triggers, highjacking of DML semantics, silent loss of data), the cure might seem worst than the disease. As far as I can tell, there is no way to reasonably emulate MERGE in SQLite, short of resorting to some external programming logic. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
