Deleting all duplicates

Question

I am trying to delete all the duplicates but keeping single record only (shorter id). Following query deletes duplicates but take lot of iterations to delete all copies and keeping original ones.

DELETE FROM emailTable WHERE id IN (
 SELECT * FROM (
    SELECT id FROM emailTable GROUP BY email HAVING ( COUNT(email) > 1 )
 ) AS q
)

Its MySQL.

DDL

CREATE TABLE `emailTable` (
 `id` mediumint(9) NOT NULL auto_increment,
 `email` varchar(200) NOT NULL default '',
 PRIMARY KEY  (`id`)
) ENGINE=MyISAM AUTO_INCREMENT=298872 DEFAULT CHARSET=latin1

Derek Downey · Accepted Answer · 2011-09-18T23:47:49.583

Try this:

DELETE FROM emailTable WHERE NOT EXISTS (
 SELECT * FROM (
    SELECT MIN(id) minID FROM emailTable    
    GROUP BY email HAVING COUNT(*) > 0
  ) AS q
  WHERE minID=id
)

The above worked for my test of 50 emails (5 different emails duplicated 10 times).

You might need to add an index on the 'email' column:

ALTER TABLE emailTable ADD INDEX ind_email (email);

It might be a bit slow fro 250,000 rows. It was slow for me on a table that had 1.5million rows (properly indexed), which is how I came up with this strategy:

/* CREATE MEMORY TABLE TO HOUSE IDs of the MIN */
CREATE TABLE email_min (minID INT, PRIMARY KEY(minID)) ENGINE=Memory;

/* INSERT THE MINIMUM IDs */
INSERT INTO email_min SELECT id FROM email
    GROUP BY email HAVING MIN(id);

/* MAKE SURE YOU HAVE RIGHT INFO */
SELECT * FROM email 
 WHERE NOT EXISTS (SELECT * FROM email_min WHERE minID=id)

/* DELETE FROM EMAIL */
DELETE FROM email 
 WHERE NOT EXISTS (SELECT * FROM email_min WHERE minID=id)

/* IF ALL IS WELL, DROP MEMORY TABLE */
DROP TABLE email_min;

The benefit of the memory table is there's an index that is used (primary key on minID) that speeds up the process over a normal temporary table.

score 4 · Answer 2 · edited Apr 13 '17 at 12:42

Here is a more streamlined deletion process:

CREATE TABLE emailUnique LIKE emailTable;
ALTER TABLE emailUnique ADD UNIQUE INDEX (email);
INSERT IGNORE INTO emailUnique SELECT * FROM emailTable;
SELECT * FROM emailUnique;
ALTER TABLE emailTable  RENAME emailTable_old;
ALTER TABLE emailUnique RENAME emailTable;
DROP TABLE emailTable_old;

Here is some sample data:

use test
DROP TABLE IF EXISTS emailTable;
CREATE TABLE `emailTable` (
 `id` mediumint(9) NOT NULL auto_increment,
 `email` varchar(200) NOT NULL default '',
 PRIMARY KEY  (`id`)
) ENGINE=MyISAM;
INSERT INTO emailTable (email) VALUES
('redwards@gmail.com'),
('redwards@gmail.com'),
('redwards@gmail.com'),
('redwards@gmail.com'),
('rolandoedwards@gmail.com'),
('rolandoedwards@gmail.com'),
('rolandoedwards@gmail.com'),
('red@gmail.com'),
('red@gmail.com'),
('red@gmail.com'),
('rolandoedwards@gmail.com'),
('rolandoedwards@gmail.com'),
('rolandoedwards@comcast.net'),
('rolandoedwards@comcast.net'),
('rolandoedwards@comcast.net');
SELECT * FROM emailTable;

I ran them. Here are the results:

mysql> use test
Database changed
mysql> DROP TABLE IF EXISTS emailTable;
Query OK, 0 rows affected (0.01 sec)

mysql> CREATE TABLE `emailTable` (
    ->  `id` mediumint(9) NOT NULL auto_increment,
    ->  `email` varchar(200) NOT NULL default '',
    ->  PRIMARY KEY  (`id`)
    -> ) ENGINE=MyISAM;
Query OK, 0 rows affected (0.05 sec)

mysql> INSERT INTO emailTable (email) VALUES
    -> ('redwards@gmail.com'),
    -> ('redwards@gmail.com'),
    -> ('redwards@gmail.com'),
    -> ('redwards@gmail.com'),
    -> ('rolandoedwards@gmail.com'),
('rolandoedwards@comcast.net');
SELECT * FROM emailTable;
    -> ('rolandoedwards@gmail.com'),
    -> ('rolandoedwards@gmail.com'),
    -> ('red@gmail.com'),
    -> ('red@gmail.com'),
    -> ('red@gmail.com'),
    -> ('rolandoedwards@gmail.com'),
    -> ('rolandoedwards@gmail.com'),
    -> ('rolandoedwards@comcast.net'),
    -> ('rolandoedwards@comcast.net'),
    -> ('rolandoedwards@comcast.net');
Query OK, 15 rows affected (0.00 sec)
Records: 15  Duplicates: 0  Warnings: 0

mysql> SELECT * FROM emailTable;
+----+----------------------------+
| id | email                      |
+----+----------------------------+
|  1 | redwards@gmail.com         |
|  2 | redwards@gmail.com         |
|  3 | redwards@gmail.com         |
|  4 | redwards@gmail.com         |
|  5 | rolandoedwards@gmail.com   |
|  6 | rolandoedwards@gmail.com   |
|  7 | rolandoedwards@gmail.com   |
|  8 | red@gmail.com              |
|  9 | red@gmail.com              |
| 10 | red@gmail.com              |
| 11 | rolandoedwards@gmail.com   |
| 12 | rolandoedwards@gmail.com   |
| 13 | rolandoedwards@comcast.net |
| 14 | rolandoedwards@comcast.net |
| 15 | rolandoedwards@comcast.net |
+----+----------------------------+
15 rows in set (0.00 sec)

mysql> CREATE TABLE emailUnique LIKE emailTable;
Query OK, 0 rows affected (0.04 sec)

mysql> ALTER TABLE emailUnique ADD UNIQUE INDEX (email);
Query OK, 0 rows affected (0.06 sec)
Records: 0  Duplicates: 0  Warnings: 0

mysql> INSERT IGNORE INTO emailUnique SELECT * FROM emailTable;
Query OK, 4 rows affected (0.01 sec)
Records: 15  Duplicates: 11  Warnings: 0

mysql> SELECT * FROM emailUnique;
+----+----------------------------+
| id | email                      |
+----+----------------------------+
|  1 | redwards@gmail.com         |
|  5 | rolandoedwards@gmail.com   |
|  8 | red@gmail.com              |
| 13 | rolandoedwards@comcast.net |
+----+----------------------------+
4 rows in set (0.00 sec)

mysql> ALTER TABLE emailTable  RENAME emailTable_old;
Query OK, 0 rows affected (0.03 sec)

mysql> ALTER TABLE emailUnique RENAME emailTable;
Query OK, 0 rows affected (0.00 sec)

mysql> DROP TABLE emailTable_old;
Query OK, 0 rows affected (0.00 sec)

mysql>

As shown the emailTable will contain the first occurrence of each email address and the corresponding original id. For this example:

IDs 1-4 have redwards@gmail.com, but only 1 was preserved.
IDs 5-7,11,12 have rolandoedwards@gmail.com, but only 5 was preserved.
IDs 8-10 have red@gmail.com, but only 8 was preserved.
IDs 13-15 have rolandoedwards@comcast.net, but only 13 was preserved.

CAVEAT : I answered a question similar to this concerning table deletion by means of a temp table approach.

Give it a Try !!!

Delux · Answer 3 · 2011-09-16T14:33:16.883

1

Here's a real quick Itzik solution. This will work in SQL 2005 and greater.

WITH Dups AS
(
  SELECT *,
    ROW_NUMBER()
      OVER(PARTITION BY email ORDER BY id) AS rn
  FROM dbo.emailTable
)
DELETE FROM Dups
WHERE rn > 1;

edited Sep 16 '11 at 14:33

answered Sep 16 '11 at 14:18

Delux

774
1
7
14

OP is asking for MySQL – Derek Downey Sep 16 '11 at 14:33
2

Yeah, just realized that; doh! Well, it's a great solution for MS SQL :) – Delux Sep 16 '11 at 14:48
Not bad to know about MS SQL too :p but at the moment looking for MySQL solution. – Gary Lindahl Sep 16 '11 at 14:54

Deleting all duplicates

3 Answers3

Linked

Related