Comments on: Oracle: Deleting Duplicate Rows Efficiently

By: true

true — Thu, 19 Jun 2008 01:49:24 +0000

Select alteranate rows

select * from Table where rowid%2=0;

By: Abhijeet

Abhijeet — Fri, 15 Sep 2006 08:49:33 +0000

Hi,

Is it possible to delete alternate rows from a table?
If yes, can some one send me the query?

Thanks
Abhijeet.

By: procter

procter — Mon, 29 May 2006 13:34:05 +0000

I like using the LEAD function, and it works well for removing duplicates, but you need to be sure that every row has exactly one duplicate. In my test scenario, it performed poorly because I had only a subset of duplicates. In such systems, you would need to narrow the results to only problem records and also factor in the possibility of one record having multiple duplicates.

By: William Robertson

William Robertson — Sat, 27 May 2006 23:19:17 +0000

Why not simply

DELETE table_name
WHERE rowid IN
( SELECT LEAD(rowid) OVER (PARTITION BY pk ORDER BY whatever)
FROM table_name );