sorry i forgot to tell you about the wait event of cbc (cache buffers chains) is within the query also. regards ujang On 10/24/07, Ujang Jaenudin <ujang.jaenudin@xxxxxxxxx> wrote: > all, > > I have varchar2 column (transc_id) with the value something like > this...and this column was indexed (unique index) > > 811120980.20070920033742 > 811120980.20071021183003 > > UPDATE c_transaction SET pgi_status = > decode(c_status,0,1,1,1,2,1,3,3), ne_status = 1, messagetracking_id = > NULL, last_processing_dtm = SYSDATE WHERE transc_id = :1 and transc_id > = :2 > > at the tkprof the cpu cost is very high , its about 30k, rows returned > is 1 , io cost is 4 , the cost itself only 4. > > the dba_indexes says: > num_rows = 2195991 > distinct_key = 1507195 > clustering_factor = 1476573 > blevel = 2 > leaf_blocks = 14486 > > histogram is already there with 254 bucket.... > > I'm guessing to turn index to reverse key, is it has the benefit impact? > > thanks in advance > > -- > regards > ujang > -- regards ujang -- //www.freelists.org/webpage/oracle-l