You haven’t mentioned a version of Oracle, and you haven’t supplied any execution plans – although we might guess from what you’ve said that the subquery is being materialized.

I’d have to guess, from what you have said, that the optimizer was using a simple filter pushdown (FPD) when the views were inline, but wasn’t able to do the pushdown when the subquery materialized. This sounds very like an (unpublished) bug that Karen Morten mentioned to me some time ago.

If you want a workaround – try creating the subquery as a real ‘in-database’ view.

Original view (V1)
————–
select *
from t1, (select a complex query here) X
where
union all
select *
from t2, (select same complex query ) X
where

subquery factoring applied to the V1 (let’s say V2)
———–
with X as
(select a complex query here)
select *
from t1, X
where
union all
select *
from t2, X
where

The report using the view V2 with a filter takes longer than the unconstrained one.
I guess, it is because it is loading all the data from the complex query into the temp table, then goes on with the rest of the query.
Is there a way to tell ORACLE to evaluate the filter first and then proceed to the temp table trans formation ?

Thanks,
Sharas

Thanks, yes, I’ve now corrected it.

is the ‘,’ behind mcc in the first query just a typo?

—————–8<—————–
…
left join
max_cust_comm mcc, — <<<<
on mcc.order_id = ord.order_id
—————–8<—————–

Shouldn't you get an ORA-00905 – Missing keyword?

So much to test, so little time ! (Sir Cecil Rhodes – misquoted).

Thanks for picking this up and letting us know the results. That’s action worthy of an Oak Table member.

----------------------------------------------------------------------------
| Id  | Operation           | Name | Rows  | Bytes | Cost (%CPU)| Time     |
----------------------------------------------------------------------------
|   0 | SELECT STATEMENT    |      |     1 |    13 |    15   (0)| 00:00:01 |
|*  1 |  FILTER             |      |       |       |            |          |
|   2 |   TABLE ACCESS FULL | T2   |     1 |    13 |     3   (0)| 00:00:01 |
|   3 |   UNION-ALL         |      |       |       |            |          |
|*  4 |    TABLE ACCESS FULL| T1   |     1 |    13 |     3   (0)| 00:00:01 |
|*  5 |    TABLE ACCESS FULL| T1   |     1 |    13 |     3   (0)| 00:00:01 |
|*  6 |    TABLE ACCESS FULL| T1   |     1 |    13 |     3   (0)| 00:00:01 |
|*  7 |    TABLE ACCESS FULL| T1   |     1 |    13 |     3   (0)| 00:00:01 |
----------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   1 - filter( EXISTS ( (SELECT "NUM" FROM "T1" "T1" WHERE "NUM"<=1)
              UNION ALL  (SELECT "NUM" FROM "T1" "T1" WHERE "NUM"<=1) UNION ALL
              (SELECT "NUM" FROM "T1" "T1" WHERE "NUM"<=1) UNION ALL  (SELECT "NUM"
              FROM "T1" "T1" WHERE "NUM"<=1)))
   4 - filter("NUM"<=1)
   5 - filter("NUM"<=1)
   6 - filter("NUM"<=1)
   7 - filter("NUM"<=1)

with explicit message in 10053 trace:

Subquery Unnesting on query block SEL$1 (#1)SU: Performing unnesting that does not require costing.
SU: Considering subquery unnest on query block SEL$1 (#1).
SU:   Checking validity of unnesting subquery SET$1 (#2)
SU:     SU bypassed: Unnest would lead to cartesian product.
SU:   Validity checks failed.

SQL> select num from t2
  2  where 1 >= any (select num as col from t1
  3             union all
  4             select num from t1
  5             union all
  6             select num from t1
  7             union all
  8             select num from t1);

                 NUM
--------------------
                   1
                   1
                   1
                   1

4 rows selected.

SQL> @x

PLAN_TABLE_OUTPUT
----------------------------------------------------------------------
----------------------------------------------------------------------
SQL_ID  11s2f4nmsbwum, child number 1
-------------------------------------
select num from t2 where 1 >= any (select num as col from t1
union all            select num from t1            union all
select num from t1            union all            select num from t1)

Plan hash value: 2572381663

---------------------------------------------------
| Id  | Operation             | Name     | E-Rows |
---------------------------------------------------
|   0 | SELECT STATEMENT      |          |        |
|   1 |  MERGE JOIN CARTESIAN |          |      1 |
|   2 |   TABLE ACCESS FULL   | T2       |      1 |
|   3 |   BUFFER SORT         |          |      4 |
|   4 |    VIEW               | VW_NSO_1 |      4 |
|   5 |     UNION-ALL         |          |        |
|*  6 |      TABLE ACCESS FULL| T1       |      1 |
|*  7 |      TABLE ACCESS FULL| T1       |      1 |
|*  8 |      TABLE ACCESS FULL| T1       |      1 |
|*  9 |      TABLE ACCESS FULL| T1       |      1 |
---------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   6 - filter("NUM"<=1)
   7 - filter("NUM"<=1)
   8 - filter("NUM"<=1)
   9 - filter("NUM"<=1)

How’s this for wrong results: http://asktom.oracle.com/pls/asktom/f?p=100:11:0::::P11_QUESTION_ID:4423923392083#2674339300346087357

You run a query with a restriction that should reduce the row count – and get back more rows than you put in !