[DERBY-6371] DROP COLUMN looks in wrong schema when checking trigger dependencies

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 10.10.1.1, 10.10.2.0
• Fix Version/s: 10.10.2.0, 10.11.1.1
• Component/s: SQL
• Issue & fix info: Repro attached
• Assignee: Knut Anders Hatlen
• Reporter: Knut Anders Hatlen
• Created: 04/Oct/13 11:52
• Updated: 21/Jan/15 00:23
• Resolved: 10/Oct/13 11:06

Description

If you define a trigger in another schema than the current schema, and the trigger action references tables in the current schema without explicit schema names, ALTER TABLE DROP COLUMN may get confused when checking dependencies.

For example:

ij> create table s1.t1(x int, y int);
0 rows inserted/updated/deleted
ij> create table s1.t2(x int, y int);
0 rows inserted/updated/deleted
ij> set schema s1;
0 rows inserted/updated/deleted
ij> create trigger s2.tr after update of x on t1 for each row insert into t2 select x from t1;
0 rows inserted/updated/deleted
ij> alter table s1.t1 drop column y restrict;
ERROR X0Y25: Operation 'DROP COLUMN' cannot be performed on object 'Y' because TRIGGER 'TR' is dependent on that object.

The ALTER TABLE statement should have succeeded, because the trigger does not depend on column Y.

If you create dummy tables in S2 with the same names as those in S1, the same ALTER TABLE statement succeeds:

ij> create table s2.t1(x int);
0 rows inserted/updated/deleted
ij> create table s2.t2(x int);
0 rows inserted/updated/deleted
ij> alter table s1.t1 drop column y restrict;
0 rows inserted/updated/deleted

The existence of these unrelated tables should not affect whether you're allowed to drop the column.

Issue Links

• is related to: DERBY-6481 Derby 10.10 backport issue (winter/spring 2014)