delete duplicate records

[vikas sharmaa]

i've an APR_MGR table with columns as mgr_id, flag, and backupid. thr is no primary key constraints in the table. the data is like below:

MgrID flag backupid
101 N null
101 N null
101 N null
102 Y 1111
103 Y 2222

now, i wud like to write a delete query, that will delete all the duplicate records of MgrID 101, except the first one.

i've written the below query:

delete from APR_MGR where MgrID = 101 and flag = 'N' and 1 < (select count(MgrID) from APR_MGR where MgrID = 101 and flag = 'N')

i was expecting 2 records to be deleted of MgrID 101, but it deleted all the 3 records.

what shall be the correct query?
[ August 30, 2008: Message edited by: vikas sharmaa ]

[Jeanne Boyarsky]

Vikas,
What database are you using? If it is one that supports the concept of a row number, you can write a query that keeps the "first" one intact.

If it does not support the concept of a row number, the procedure is more steps. You will have to copy the distinct records into a new table, delete all the records from this table and then copy the records back into this table.

[vikas sharmaa]

i m using oracle 9i. then what shall be the delete query?

[Avi Abrami]

Vikas,
This has nothing to do with JDBC of-course :-)
(NOTE: Untested - and check the syntax also :-)

create table APR_MGR_TMP as
select distinct MGRID, FLAG, BACKUPID from APR_MGR
/
drop table APR_MGR
/
rename APR_MGR_TMP as APR_MGR
/

Good Luck,
Avi.

[Jeanne Boyarsky]

Vikas,
It's easier to do this in Oracle than some other databases so this is good. See this wiki page sql for an example.

It's ok the question isn't about JDBC. It is about SQL making it on topic.