Effective solution for algorithm

Question

I have table T defined below.
1. Table t definition.
with t as
(select 'A' CODE from dual
union all
select 'A' CODE from dual
union all
select 'B' from dual)
select * from t;1) In plsql code i have to return cursor ("o_cur") over table t, that i can achieve as:
open o_cur for
select * from t;
2) Also input varivale "i_code" is given and an output variable "o_code" should be returned with following logic:
if o_cur has in its records value "i_code" then return o_code = i_code,
else if, o_cur has exactly on row then return o_code value as the value in that one record,
else return o_code=null;
What is the effecients way to achieve step 2?
I can do un-effeciently those 2 queries:
2.1) select count(1)
into v_count from t;
2.2) select count(1)
into v_count from t
where CODE = i_code;
And then wrote the logic in step "2)" to get output variable o_code value. But this seems slow.
Maybe i can query already opened cursor "o_cur" somehow? Or is there better way?
I only can do this way:
with t as
(select 'A' CODE from dual
union all
select 'A' CODE from dual
union all
select 'B' from dual)
select count(1) CNT
, max((case when t.CODE = 'A' then 'Y' else 'N' end)) YN
from t ;--rowcount=3, 'A' in list
with t as
(select 'A' CODE from dual
union all
select 'A' CODE from dual
union all
select 'B' from dual)
select count(1) CNT
, max((case when t.CODE = 'C' then 'Y' else 'N' end)) YN
from t ;--rowcount=3, 'A' is not in listSeems it is the best solution.
Edited by: CharlesRoos on Apr 6, 2010 3:57 AM

Answer 1

CharlesRoos wrote:
Thx,
with "count(*) over () as total_records"
i can add to output cursor additional column, which have on all rows same value showing the count of all records. So if i add this additional column to output cursor "o_cur" then i can fetch one row from it, close cursor, and open again. And fro mfetching the one record i get information about the total count of records. But as you see algorithm i nstep "2)" also requires knowledge of weather the input value "i_code" appeared in cursor resultset.
I think i better do additional query after openeing the cursor like:
with t as
(select 'A' CODE from dual
union all
select 'A' CODE from dual
union all
select 'B' from dual)
select count(1) CNT
, max((case when t.CODE = 'C' then 'Y' else 'N' end)) YN
from t ;--rowcount=3, 'C' is not in list
Another analytical function can be used to test for the existence of a value in the result set, e.g.
SQL> ed
Wrote file afiedt.buf
  1  select empno, ename, job, sal
  2        ,row_number() over (order by empno) as rn
  3        ,count(*) over () as cnt
  4        ,sum(decode(job,'CLERK',1)) over () as clerks
  5  from emp
  6* where deptno = 10
SQL> /
     EMPNO ENAME      JOB              SAL         RN        CNT     CLERKS
      7782 CLARK      MANAGER         2450          1          3          1
      7839 KING       PRESIDENT       5000          2          3          1
      7934 MILLER     CLERK           1300          3          3          1
SQL>Here I've used the sum function to sum up the number of CLERK jobs in the results. This shows a "1" for all rows indicating there is 1 CLERK in the result set, even if it hasn't yet been selected by the cursor.