SQL & PL/SQL :: Replicating Records With Cartesian Join

Table script:

CREATE TABLE TEST_ITEM_BU_ID
(
CCN VARCHAR2(100 CHAR),
SKU VARCHAR2(100 CHAR),
BU_ID NUMBER
)

select * from test_item_bu_id;

CCN SKU BU_ID
------------------------------
M10000616-10502414545
M10000600-11437414545
M10000205-113380
M10000205-113390
M10000600-114370

The requirement is to replicate the bu_id records with bu_id=0 as bu_id=414545 ( there is no lookup available) so the same table should act as a lookup table to populate bu_id for the records where bu_id=0

i.e., here it will replicate for the sku set with bu_id value=0

M10000205-113380
M10000205-113390
M10000600-114370

will be replicated against

M10000616-10502414545M10000600-11437414545

so the output should be :

CCN SKU BU_ID
------------------------------
M10000205-11338414545
M10000205-11338414545
M10000205-11339414545
M10000205-11339414545

The below query is supposed to do this.

select a.ccn,b.bu_id,a.sku,b.sku
from test_item_bu_id a ,
( select distinct ccn,sku_num, bu_id
from test_item_bu_id
where bu_id in (414545) and CCN in ('M10000') ) b
where a.bu_id = 0 and a.sku <> b.sku and a.ccn= b.ccn

But we have wrong result here.

CCN BU_ID SKU SKU_1
----------------------------------------------
M10000414545205-11338600-11437
M10000414545205-11338616-10502
M10000414545205-11339600-11437
M10000414545205-11339616-10502
M10000414545600-11437616-10502

How can we avoid the last record, i.e., SKU=600-11437 since it is already having bu_id no need to replicate it, but it is getting replicated since the extra record with bu_id=0 exist for the same sku.