Can someone explain me that unexpected result with dbms_random ?

Question

Hi There,
I wonder how executing dbms_random only once could return two different values?
select x,x from (select dbms_random.value(0,100) x from dual)
         X          X
62,6555943 31,0639443Thanks a lot for your feedback
Laurent

Answer 1

The optimizer the merged view into main query and called the function twice (whether or not it should be allowed to do this with non-deterministic function is another matter). Compare explain plan and results below with those with NO_MERGE or ROWNUM to prevent merging of the view.
Oracle9i Enterprise Edition Release 9.2.0.5.0 - 64bit Production
With the Partitioning, OLAP and Oracle Data Mining options
JServer Release 9.2.0.5.0 - Production
SQL> SET AUTOTRACE ON EXPLAIN
SQL> /* Original */
SQL> SELECT x, x
  2  FROM (
  3    SELECT DBMS_RANDOM.VALUE (0, 100) x
  4    FROM   dual);
         X          X
89.1805061 93.2011152
Execution Plan
   0      SELECT STATEMENT Optimizer=CHOOSE
   1    0   TABLE ACCESS (FULL) OF 'DUAL'
SQL> /* With NO_MERGE hint in view */
SQL> SELECT x, x
  2  FROM (
  3    SELECT /*+ NO_MERGE */ DBMS_RANDOM.VALUE (0, 100) x
  4    FROM   dual);
         X          X
39.9149281 39.9149281
Execution Plan
   0      SELECT STATEMENT Optimizer=CHOOSE (Cost=17 Card=8168 Bytes=1
          06184)
   1    0   VIEW (Cost=17 Card=8168 Bytes=106184)
   2    1     TABLE ACCESS (FULL) OF 'DUAL' (Cost=17 Card=8168)
SQL> /* With ROWNUM in view */ 
SQL> SELECT x, x
  2  FROM (
  3    SELECT DBMS_RANDOM.VALUE (0, 100) x, ROWNUM
  4    FROM   dual);
         X          X
35.2401179 35.2401179
Execution Plan
   0      SELECT STATEMENT Optimizer=CHOOSE
   1    0   VIEW
   2    1     COUNT
   3    2       TABLE ACCESS (FULL) OF 'DUAL'
SQL>