A very slow select with sub selects sql statement on Oracle

Question

Hi,
I'm moving an application from MySql to Oracle. The following select were very efficiently executed in MySql. In oracle its slow like a snail.
Do anyone have a hint on how to speed it up?
The slow part is the four sub selects in the select part. Removing them makes the select about 50 times faster on Oracle.
Best Regards,
Stephane
select
(select count(*) from relation rr where rr.document_id = d.id and rr.product_id = ? and (rr.relation_type_id = 'link' OR rr.relation_type_id = 'product')) as relationList,
(select count(*) from relation rr where rr.document_id = d.id and rr.product_id = ? and rr.relation_type_id = 'number') as relationNumber,
(select count(*) from relation rr where rr.document_id = d.id and rr.product_id = ? and rr.relation_type_id = 'title') as relationTitle,
(select count(*) from relation rr where rr.document_id = d.id and rr.product_id = ? and rr.relation_type_id = 'content') as relationText,
d.*
from document d,(
select distinct r.document_id id
from relation r
where
r.product_id = ?
) dd
where d.id=dd.id

Answer 1

You are accessing the relation-table too many times, so a rewrite to a query like this
SQL> select dept.deptno
  2       , dept.dname
  3       , count(decode(job,'CLERK',1)) clerk
  4       , count(decode(job,'MANAGER',1)) manager
  5       , count(decode(job,'SALESMAN',1)) salesman
  6    from dept, emp
  7   where dept.deptno = emp.deptno (+)
  8   group by dept.deptno
  9       , dept.dname
10  /
    DEPTNO DNAME               CLERK    MANAGER   SALESMAN
        10 ACCOUNTING              1          1          0
        20 RESEARCH                2          1          0
        30 SALES                   1          1          4
        40 OPERATIONS              0          0          0
4 rijen zijn geselecteerd.will be worth the effort.
If still not satisfied, you have to do some investigation, as described [url http://forums.oracle.com/forums/thread.jspa?threadID=501834&tstart=0]here
Regards,
Rob.