MikeW <[EMAIL PROTECTED]> wrote: > I presume I'm not allowed to do this: > > // apply updates on first db - start Transaction > ret = sqlite3_exec( > first_db, > "BEGIN; UPDATE " FIRST_DB_TABLE > " SET " FIRST_DB_COLUMN "=1 " > " WHERE " FIRST_DB_COLUMN " NOT NULL;", > NULL, NULL, NULL)); > // returns SQLITE_OK > > // delete updates from second db - commit Transaction > ret = sqlite3_exec( > second_db, > "UPDATE " SECOND_DB_TABLE > " SET " SECOND_DB_COLUMN "= NULL " > " WHERE " SECOND_DB_COLUMN " NOT NULL; END;", > NULL, NULL, NULL)); > // returns SQLITE_INTERNAL
Do the two DB connection handles (first_db and second_db) refer to the same file? However, in this case I would expect SQLITE_BUSY, not SQLITE_INTERNAL. There must be some other problem, in the code you don't show. > Still, I would like to be able to run the Transaction over these two > calls You can't. A transaction is a property of a connection. You can't have a single transaction spanning multiple connections. You can, however, have the same connection refer to more than one database file. See ATTACH statement. This way, you can update several files within a single transaction. > ... What's the best way, given I have two separate handles ? > Make them separate Transactions and do a ROLLBACK on the first if the > second exec fails, possibly ? That's unreliable. E.g., you run the first transaction but keep it open (so you can roll it back). You run the second transaction - it succeeds. But before you have a chance to commit the first one, there's a power failure. When your program is restarted, the first transaction will be rolled back, and you will end up with an inconsistent state. Igor Tandetnik _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
