Oracle deleting duplicates based on a condition

问题

I am trying to delete some duplicates out of a table and trying to retain one of the duplicates for ID1 column.

Using this query I managed to delete based on ROWID.

delete from    tabela.lorik
where ROWID not in (
select MAX(ROWID) 
from     tabela.lorik
GROUP BY ID1)

Now I want to delete all those records found duplicate where NETAMT = 0

回答1:

You can achieve this using the following query:

delete from  tabela.lorik O
where O.netamt = 0 
AND EXISTS (SELECT 1 FROM tabela.lorik I
WHERE I.ID = O.ID AND I.netamt <> 0)

I am assuming that you need to delete only records where it is netamount = 0, If not then do comment below.

If you are looking to retain one non-zero entry and delete all others(In case of all the zeros, one entry with zero as netamount will be retained) then you can use the following query:

DELETE FROM TABELA.LORIK O
WHERE
    ROWID IN (
        SELECT
            RWID
        FROM
            (
                SELECT
                    ROWID AS RWID,
                    ROW_NUMBER() OVER(
                        PARTITION BY ID
                        ORDER BY
                            CASE
                                WHEN NETAMT = 0 THEN 2
                                ELSE 1
                            END
                    ) AS RN
                FROM
                    TABELA.LORIK
            )
        WHERE
            RN > 1
    );

Cheers!!

回答2:

You can order the rowids by the value first and then rowid. This following keeps only one row, with a preference for the most recent non-zero amount:

delete from tabela.lorik
where ROWID <> (select max(rowid) keep (dense_rank first (order by (case when amount = 0 then 1 else 2 end), rowid desc)
                from tabela.lorik l2
                where l2.id1 = l.id1
               )

来源：https://stackoverflow.com/questions/58485689/oracle-deleting-duplicates-based-on-a-condition

标签

sql

Oracle

duplicates