Question

I need to find and delete duplicates from within a table, yet keep a record of the IDs of the records being deleted and add them to a ‘deleted_ids’ field.

For example, here’s part of a table I have currently:

+--------+-------------------------------------------------------+
| Id     | Name                                                  |
+--------+-------------------------------------------------------+
|   9627 |  John Smith                                           |
|   9657 |  John Smith                                           |
|  24455 |  John Smith                                           |
|  67435 |  John Smith                                           |
+--------+-------------------------------------------------------+

The result I’d like to achieve is:

+--------+-----------------------------+-------------------------+
| Id     | Name                        | Deleted_Ids             |
+--------+-----------------------------+-------------------------+
|   9627 |  John Smith                 | 9657,24455,67435        |
+--------+-----------------------------+-------------------------+

I’m open to any practical approach, if anyone can provide a solution or give suggestions I’d really appreciate it.

Thanks,
Paul

Answer 1

If you want to proceed like this, here is a simple way.
Do it all in one transaction to be sure.
If there could be an entry of 'John Smith' in tbl_del already, you have to check and UPDATE to add the deleted ids instead.

BEGIN;

INSERT INTO tbl_del
SELECT name, GROUP_CONCAT(id) AS deleted
FROM   tbl_live
WHERE  name = 'John Smith'
GROUP  BY 1;

DELETE FROM tbl_live
WHERE  name = 'John Smith';

COMMIT;

But I think Veseliq has a vaid point.