MySQL duplicate entry deletion in 5.5.9

Question

Hey all,
I've noticed that in the newest MySQL, I can no longer do:
ALTER TABLE test ADD UNIQUE INDEX unique_index (id);
ALTER TABLE test DROP INDEX unique_index;
to remove duplicates. I get the following error:
ERROR 1062 (23000): Duplicate entry '3' for key 'unique_index']
This worked fine before the major SQL upgrade. It also works fine on version 5.1.52 on my web host.  I searched through the 5.5.9 release notes and I didn't see anything that should have changed this, but I'm no MySQL guru! Is this a feature or a bug? Any alternate suggestions for deleting duplicates that will work on this version and older versions of MySQL?
Thanks!
Scott
Edit: Just for clarity, the table I'm working on has a single primary key column (UNIQUE_KEY) (with 80+ other data columns) and I'm only detecting the duplicate row based on the value of the primary key column.
Last edited by firecat53 (2011-02-25 03:33:30)

Answer 1

Alright, since I was apparently sucked in by this particular method the first time I googled the answer, care to enlighten me on the 'more practical' non-joke methods, as there appears to be more than one option.
Also, why is that a joke of a feature? Like I said, I'm no MySQL guru so I'm ready to learn!
Thanks,
Scott
edit: is this a 'more practical' method? It's hard when you're learning on your own to sort out the good from the bad sometimes
edit2: using the above method with this table:
+------+------+-------+
| id | data | data1 |
+------+------+-------+
| 1 | 10 | 100 |
| 2 | 100 | 1000 |
| 3 | 1000 | 500 |
| 3 | 1000 | 500 |
+------+------+-------+
and this query:
delete test from test left outer join (select min(id) as RowID, id from test group by id) as KeepRows on test.id=KeepRows.RowID where KeepRows.RowID is null;
I can't quite get it to work...it doesn't delete any rows. What am I missing there?
edit3: This method works -- is it a 'good' method or should I be looking at something different? Again, I'm just matching on duplicates in the primary key column (before the primary key is assigned):
create table test1 like test;
insert into test1 select * from test group by id;
drop table test;
rename table test1 to test;
Last edited by firecat53 (2011-02-25 03:55:10)