Comments on: How to delete duplicate rows with SQL, Part 2

By: Diwakar

Diwakar — Mon, 12 Nov 2007 13:36:11 +0000

BINGO !!!!
awesome article

CHEERS!!
—
Diwakar

By: Joseph

Joseph — Wed, 17 Oct 2007 10:34:21 +0000

i want to delete the records on my table that has simllar entries in my column. e.g column A: has the following entries [aaaaaaa]
[aaaaamobile]
[aaaamotoraaaa]
now i want to delete only the a’s

By: jimro

jimro — Sat, 03 Mar 2007 12:26:58 +0000

Hi, sorry for my poor english.

I forgot to format a part of my previous message in XHTML, thank you for correction.

With this least(0, length(@type := type))) > 0 only one row is preserved. Isn't this the required goal?

Regarding my previous code, it would be better like this:

set @num := 0, @source := '', @target := '';
delete from test
where greatest(0,
   @num := if(source = @source and target = @target, @num + 1, 0),
   least(0, length(@source := source), length(@target := target))) > 0
order by source, target;

By: jimro

jimro — Fri, 02 Mar 2007 22:40:20 +0000

Hi, thanks for these nice articles.

Maybe least(0, length(@type := type))) > 0 instead of least(0, length(@type := type))) > 1

I've tested with two columns (MySQL 5.0.27) and it works very well

set @num := 0, @source := '', @target := '';
delete from test
where greatest(0,
   @num := if(source = @source and target = @target, @num + 1, 0),
   least(0, length(@source := source), length(@target := target))) > 0
order by source;

By: Collin

Collin — Fri, 09 Feb 2007 01:59:04 +0000

I have a genius method for removing duplicates.. open enterprise manager.. list all rows and then go through deleting the rows that are the same using the delete button.

OK, OK.. not so JENIUS after all.. worked for me though.

By: Xaprb

Xaprb — Wed, 07 Feb 2007 12:51:06 +0000

Thanks Shantanu, I didn’t know about that!

By: Shantanu Oak

Shantanu Oak — Wed, 07 Feb 2007 12:08:53 +0000

Why not to add unique index to drop duplicate rows? It works only in MySQL.
alter ignore table dupTest add unique index (a,b)

By: Xaprb

Xaprb — Tue, 06 Feb 2007 15:17:58 +0000

Bill Minton: bravo! Thanks for writing this in. The SET ROWCOUNT is exactly the type of platform-specific trick I was talking about.

By: Bill Minton

Bill Minton — Tue, 06 Feb 2007 15:11:41 +0000

SQL Server Version:

set nocount on
create table #tbFruit (Fruit varchar(20))
insert into #tbFruit (Fruit) values ('apple')
insert into #tbFruit (Fruit) values ('apple')
insert into #tbFruit (Fruit) values ('apple')
insert into #tbFruit (Fruit) values ('apple')
insert into #tbFruit (Fruit) values ('orange')
insert into #tbFruit (Fruit) values ('orange')
insert into #tbFruit (Fruit) values ('orange')

create table #tbFruitsWithDupes (DupeID int identity, Fruit varchar(20), DupeCount int)

insert into #tbFruitsWithDupes (Fruit, DupeCount)
	select Fruit, count(*) DupeCount from #tbFruit
	group by Fruit having count(*) > 1

declare @x int, @max int, @Fruit varchar(20), @DupeCount int

select @x = 1, @max	= max(DupeID) from #tbFruitsWithDupes

while (@x <= @max)
	begin
		select @Fruit = Fruit, @DupeCount = DupeCount - 1
			from #tbFruitsWithDupes Where DupeID = @x

		set rowcount @DupeCount

		delete from #tbFruit where Fruit = @Fruit

		select @x = @x + 1
	end

select * from #tbFruit

drop table #tbFruit
drop table #tbFruitsWithDupes