I definitely second this since I actually got bitten by this last week :-) We split over 2 databases for the same reason as Derrell and always attach after opening them. Surely it's harmless to remove these checks - if something goes wrong, the trigger will just fail normally with a "no such table"?
>> Should I remove the tests from SQLite that prevent >> triggers in one database from referring to tables in >> a different database? Or should I leave things as >> they are and close ticket #1689 with a remark of >> "works as designed". > I have had use for triggers that could operate on tables in a different > database, and been unable to use them. I separate tables into two primary > databases because the tables in one of them can be locked for long periods of > time (long transactions) while the tables in the other database are accessed > frequently. The mechanism for determining which databases will be operated on > is external to SQLite, so I would know that it is safe to issue a request that > would cause a trigger to operate on the other database. > I'd vote to allow a trigger to operate on a table in a different database, if > there are no reasons other than "they shouldn't need to do that" preventing > it. > Derrell -- Best regards, Taka mailto:[EMAIL PROTECTED]
