Rank by

Question

Hi all,
Can anybody help me........ i am using
"rank() over (partition by orh.orh_id order by ol.orl_id asc) as seq_num " this is a part of my inner query.... The whole query is using a left outer join.... Now the problem is that... when I am selecting ORH_ID and SEQ_NUM using a cursor and inserting into a table... in which ORH_ID and SEQ_NUM are PKs.... I noticed some rejections due to unique key constraint.... That means I am fetching some duplicate ORH_IDs and SEQ_NUM... How is it possible to have duplicates of ORH_ID and SEQ_NUM together... because for each occurence of same ORH_ID will have a different sequence number based on ORL_ID....
If anybody has any idea to solve this please reply... Thanks in advance...
Message was edited by:
user626688

Answer 1

Possible, when you have duplicate records of orh_id and orl_id combination.
See this.
SQL> with t as
  2  (select 1 col1, 10 col2 from dual
  3  union all
  4  select 2, 20 from dual
  5  union all
  6  select 1, 30 from dual
  7  union all
  8  select 1, 10 from dual
  9  )
10  select col1, col2 from t
11  order by 1, 2
12  /
      COL1       COL2
         1         10
         1         10
         1         30
         2         20
SQL> with t as
  2  (select 1 col1, 10 col2 from dual
  3  union all
  4  select 2, 20 from dual
  5  union all
  6  select 1, 30 from dual
  7  union all
  8  select 1, 10 from dual
  9  )
10  select col1, rank() over (partition by col1 order by col2) rn
11  from t
12  /
      COL1         RN
         1          1
         1          1
         1          3
         2          1
SQL>Cheers
Sarma.