New answers tagged jdbc
1
This could happen if:
- Your table MYTABLE has a unique column UNIQCOL
- This MYTABLE.UNIQCOL is referenced by some column in another table, say MYTABLE2.UNIQCOL_REF
- This MYTABLE2.UNIQCOL_REF is not indexed.
Adding a (non-unique) index to MYTABLE2.UNIQCOL_REF could then solve the problem.
(You said all FKs in MYTABLE are indexed, but you didn't say ...
2
Consider a parent/child table such as client/order.
You can't delete a client that has an order. Say client 123 has an order A123. Fred does a delete for that order but does not commit. Then "Jane" tries to delete client 123.
Since Fred's statement can potentially rollback, the client can't be deleted because it isn't allowed to the leave the order ...
0
i would check that the deltee is using the index. run explain plan and make sure the execution plan includes a unique index access path. this is trivial but god is in the details.
maybe what you got here is a hot spot issue. are the rows you are deleting/inserting are of high consecutive pk values ?
b-tree index is stored as a sorted index - entries are ...
2
Delete is a DML command and stores the data in redo log till the delete operation is committed.
This means that if data to be removed by delete is slightly large[even though search time is less] it will take longer time as it will move data to redo log.
So may be the instance when your operation took longer large no. of rows were being deleted to many ...
Top 50 recent answers are included
