Insert or Not
-
Got sniped at 15:30 with a load-testing environment that was losing 6% of data along the way. Got isolated to a database insert: 100% of rows was going into the INSERT statement, only 94% ended up in the table. How much of it ends up in the DB seems to vary with insert batch size.
I went home anyway but it feels like it's gonna be one of those weekends.
-
Got sniped at 15:30 with a load-testing environment that was losing 6% of data along the way. Got isolated to a database insert: 100% of rows was going into the INSERT statement, only 94% ended up in the table. How much of it ends up in the DB seems to vary with insert batch size.
How does a database mangle an INSERT that badly? (Also, just who made such a terrible excuse for a DBMS?)
-
How does a database mangle an INSERT that badly?
Triggers, is all I can think of.
(Also, just who made such a terrible excuse for a DBMS?)
Don't blame the DBMS for running triggers it's told to run.
-
How does a database mangle an INSERT that badly? (Also, just who made such a terrible excuse for a DBMS?)
Oracle. Who else.
Triggers, is all I can think of.
Plain insert, up to 16k rows per execute, no triggers whatsoever. One index.
-
-
Triggers, is all I can think of.
I would be quite astonished if the amount of rows a trigger deleted varied with the insert batch size...
Plain insert, up to 16k rows per execute, no triggers whatsoever. One index.
and yeah. HOWTF is Orrible Belgium-ing an INSERT that badly?
Filed under: Join the Larry Ellison Coal Shipment Fund today!
-
and yeah. HOWTF is Orrible ■■■■■■■-ing an INSERT that badly?
Tell me when you find out...
-
genetic algorithms and natural selection?
i'm sure that 6% could'n adapt to the medium
-
Also occurs on the old hardware (and older OS version) after updating the database... so probably a DB fault.
Guess that version's going on the blacklist.