On Thu, Jun 27, 2013 at 6:05 AM, Patrik Nilsson <nipatriknils...@gmail.com>wrote:
> Hi All! > > A feature I'm missing is a syntax like with "insert or update". > > You define a table as: > "create table table_test (id as integer primary key, a as integer, b as > integer, c as integer)" > > Then you know that the "id" is unique and you only can have one row with > that integer. > > Then you can give the following statements: > insert or update into table_test (id,c) values (1,3) > REPLACE INTO table_test SELECT 1, a, b, 3 FROM table_test WHERE id=1 UNION ALL SELECT 1, NULL, NULL, 3 LIMIT 1; > insert or update into table_test (id,b) values (1,2) > insert or update into table_test (id,a) values (1,1) > insert or update into table_test (id,a) values (5,13) > > This result is the following set: > 1|1|2|3 > 5|13|| > > Now I'm doing: "insert or ignore into table_test (id) values (1)" and > then issue an update statement. I think "insert or update" will increase > performance. > > If the "insert or update" can't perform its operation, it can issue a > SQLITE_AMBIGUOUS error. > > Best regards, > Patrik > > -- > ASCII ribbon campaign ( ) > against HTML e-mail X > www.asciiribbon.org / \ > _______________________________________________ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users > -- D. Richard Hipp d...@sqlite.org _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
