Multi view ( table & Graph ) answer takes long time

Question

Hi,
we are using OBIEE 10g and a user has created a report that uses both a table and a graph for representing the results.  When i use either of them ( on its own) the response of the system is pretty good ( about 30-40 seconds). The problem is that If i use both the response increases up to 3,5 minutes. I would like to ask if using both the table an the graph forces the engine to make the fetches twice. If so, is there a way to change this behavior, maybe through a parameter option that i am not aware of ?  ( fetch data once and then create all the presentations : table, graph).
Here are the queries and their respective execution plans that are used by the optimizer, in case they are of any help.  My apologies if i am posting them at the wrong forum
Thank you in advance
Theodore
1a.Simple query (using either a table or a graph)
/* Formatted on 6/12/2013 7:53:27 πμ (QP5 v5.139.911.3011) */
  SELECT DISTINCT D1.c5 AS c1,
                  D1.c6 AS c2,
                  D1.c4 AS c3,
                  D1.c1 AS c4,
                  D1.c3 AS c5,
                  D1.c3 / NULLIF (D1.c1, 0) AS c6,
                  D1.c2 AS c7,
                  D1.c2 / NULLIF (D1.c1, 0) AS c8
    FROM (  SELECT                                   /*+ PARALLEL(T16913,8) */
                  SUM (T16913.PHA_QUANTITY) AS c1,
                   SUM (T16913.EXEC_COST) AS c2,
                   SUM (T16913.EXEC_VALUE) AS c3,
                   COUNT (DISTINCT T366.BARCODE) AS c4,
                   T19403.YEAR_MONTH AS c5,
                   T7252.PERCENTAGE AS c6
              FROM EXECALENDAR_DIM T19403,
                   DRUG_DIM T366,
                   PERCENTAGE_DIM T7252,
                   PRESCDRUG_FACT T16913
             WHERE (    T366.DWHKEY = T16913.DRU_DWHKEY
                    AND T7252.DWHKEY = T16913.PER_DWHKEY
                    AND T16913.EXECUTION_DATE = T19403.CALENDAR_DATE
                    AND T19403.YEAR_MONTH = '201301')
          GROUP BY T7252.PERCENTAGE, T19403.YEAR_MONTH) D1
ORDER BY c1, c2
1b. Execution plan for (1a)
ID
PID
Operation
Name
Rows
Bytes
Cost
CPU Cost
IO Cost
Temp space
IN-OUT
PQ Dist
PStart
PStop
0
SELECT STATEMENT
4
252
4279
3G
4176
1
0
  PX COORDINATOR
2
1
    PX SEND QC (ORDER)
SYS.:TQ10006
4
252
4279
3G
4176
P->S
QC (ORDER)
3
2
      SORT ORDER BY
4
252
4279
3G
4176
PCWP
4
3
        PX RECEIVE
4
252
4279
3G
4176
PCWP
5
4
          PX SEND RANGE
SYS.:TQ10005
4
252
4279
3G
4176
P->P
RANGE
6
5
            SORT GROUP BY
4
252
4279
3G
4176
PCWP
7
6
              PX RECEIVE
4
252
4279
3G
4176
PCWP
8
7
                PX SEND HASH
SYS.:TQ10004
4
252
4279
3G
4176
P->P
HASH
9
8
                  SORT GROUP BY
4
252
4279
3G
4176
PCWP
10
9
                    PX RECEIVE
4
252
4279
3G
4176
PCWP
11
10
                      PX SEND HASH
SYS.:TQ10003
4
252
4279
3G
4176
P->P
HASH
12
11
                        SORT GROUP BY
4
252
4279
3G
4176
PCWP
13
12
                          HASH JOIN
10M
601M
4199
745M
4176
PCWP
14
13
                            BUFFER SORT
PCWC
15
14
                              PX RECEIVE
31820
466K
199
11M
199
PCWP
16
15
                                PX SEND BROADCAST
SYS.:TQ10000
31820
466K
199
11M
199
S->P
BROADCAST
17
16
                                  TABLE ACCESS FULL
DRUG_DIM
31820
466K
199
11M
199
18
13
                            HASH JOIN
10M
457M
3995
592M
3977
PCWP
19
18
                              BUFFER SORT
PCWC
20
19
                                PX RECEIVE
5
30
2
8071
2
PCWP
21
20
                                  PX SEND BROADCAST
SYS.:TQ10001
5
30
2
8071
2
S->P
BROADCAST
22
21
                                    TABLE ACCESS FULL
PERCENTAGE_DIM
5
30
2
8071
2
23
18
                              NESTED LOOPS
10M
400M
3989
450M
3975
PCWP
24
23
                                BUFFER SORT
PCWC
25
24
                                  PX RECEIVE
PCWP
26
25
                                    PX SEND BROADCAST
SYS.:TQ10002
S->P
BROADCAST
27
26
                                      TABLE ACCESS FULL
EXECALENDAR_DIM
30
390
29
1953824
29
28
23
                                PX BLOCK ITERATOR
328K
9309K
132
14M
132
PCWC
KEY
KEY
29
28
                                  TABLE ACCESS FULL
PRESCDRUG_FACT
328K
9309K
132
14M
132
PCWP
KEY
KEY
2a. Complicated query (using both  a table or a graph)
/* Formatted on 6/12/2013 7:34:07 πμ (QP5 v5.139.911.3011) */
  SELECT DISTINCT D1.c4 AS c1,
                  D1.c5 AS c2,
                  D1.c6 AS c3,
                  D1.c7 AS c4,
                  D1.c8 AS c5,
                  D1.c9 AS c6,
                  D1.c10 AS c7,
                  D1.c11 AS c8,
                  D2.c12 AS c9,
                  D2.c13 AS c10,
                  D2.c14 AS c11,
                  D1.c1 AS c12,
                  D1.c2 AS c13,
                  D1.c3 AS c14
    FROM (  SELECT MAX (D1.c4) AS c1,
                   MAX (D1.c3) / NULLIF (MAX (D1.c1), 0) AS c2,
                   MAX (D1.c2) / NULLIF (MAX (D1.c1), 0) AS c3,
                   D1.c10 AS c4,
                   D1.c5 AS c5,
                   D1.c9 AS c6,
                   D1.c6 AS c7,
                   D1.c8 AS c8,
                   D1.c8 / NULLIF (D1.c6, 0) AS c9,
                   D1.c7 AS c10,
                   D1.c7 / NULLIF (D1.c6, 0) AS c11
              FROM (SELECT SUM (D1.c6) OVER (PARTITION BY D1.c5) AS c1,
                           SUM (D1.c7) OVER (PARTITION BY D1.c5) AS c2,
                           SUM (D1.c8) OVER (PARTITION BY D1.c5) AS c3,
                           SUM (D1.c4) OVER (PARTITION BY D1.c5) AS c4,
                           D1.c5 AS c5,
                           D1.c6 AS c6,
                           D1.c7 AS c7,
                           D1.c8 AS c8,
                           D1.c9 AS c9,
                           D1.c10 AS c10
                      FROM (  SELECT COUNT (
                                        CASE D1.c12
                                           WHEN 1 THEN D1.c11
                                           ELSE NULL
                                        END)
                                        AS c4,
                                     D1.c5 AS c5,
                                     SUM (D1.c13) AS c6,
                                     SUM (D1.c14) AS c7,
                                     SUM (D1.c15) AS c8,
                                     COUNT (DISTINCT D1.c11) AS c9,
                                     D1.c10 AS c10
                                FROM (SELECT         /*+ PARALLEL(T16913,8) */
                                            T7252.
                                             PERCENTAGE
                                                AS c5,
                                             T19403.YEAR_MONTH AS c10,
                                             T366.BARCODE AS c11,
                                             ROW_NUMBER ()
                                             OVER (
                                                PARTITION BY T7252.PERCENTAGE,
                                                             T366.BARCODE
                                                ORDER BY
                                                   T7252.PERCENTAGE DESC,
                                                   T366.BARCODE DESC)
                                                AS c12,
                                             T16913.PHA_QUANTITY AS c13,
                                             T16913.EXEC_COST AS c14,
                                             T16913.EXEC_VALUE AS c15
                                        FROM EXECALENDAR_DIM T19403,
                                             DRUG_DIM T366,
                                             PERCENTAGE_DIM T7252,
                                             PRESCDRUG_FACT T16913
                                       WHERE (T366.DWHKEY = T16913.DRU_DWHKEY
                                              AND T7252.DWHKEY = T16913.PER_DWHKEY
                                              AND T16913.EXECUTION_DATE =
                                                     T19403.CALENDAR_DATE
                                              AND T19403.YEAR_MONTH = '201305')) D1
                            GROUP BY D1.c5, D1.c10) D1) D1,
                   (SELECT                           /*+ PARALLEL(T16913,8) */
                          SUM (T16913.PHA_QUANTITY)
                              AS c1,
                           SUM (T16913.EXEC_COST) AS c2,
                           SUM (T16913.EXEC_VALUE) AS c3,
                           COUNT (DISTINCT T366.BARCODE) AS c4
                      FROM EXECALENDAR_DIM T19403,
                           DRUG_DIM T366,
                           PRESCDRUG_FACT T16913
                     WHERE (    T366.DWHKEY = T16913.DRU_DWHKEY
                            AND T16913.EXECUTION_DATE = T19403.CALENDAR_DATE
                            AND T19403.YEAR_MONTH = '201305')) D2
          GROUP BY D1.c7 / NULLIF (D1.c6, 0),
                   D1.c8 / NULLIF (D1.c6, 0),
                   D1.c5,
                   D1.c6,
                   D1.c7,
                   D1.c8,
                   D1.c9,
                   D1.c10) D1,
         (SELECT MAX (D2.c4) AS c12,
                 MAX (D2.c3) / NULLIF (MAX (D2.c1), 0) AS c13,
                 MAX (D2.c2) / NULLIF (MAX (D2.c1), 0) AS c14
            FROM (SELECT SUM (D1.c6) OVER (PARTITION BY D1.c5) AS c1,
                         SUM (D1.c7) OVER (PARTITION BY D1.c5) AS c2,
                         SUM (D1.c8) OVER (PARTITION BY D1.c5) AS c3,
                         SUM (D1.c4) OVER (PARTITION BY D1.c5) AS c4,
                         D1.c5 AS c5,
                         D1.c6 AS c6,
                         D1.c7 AS c7,
                         D1.c8 AS c8,
                         D1.c9 AS c9,
            

Answer 1

select count(table2_pk) from view where table2_pk=1234
returns in ~ 1 second.
The plan is
SELECT STATEMENT
NESTED LOOPS (OUTER)
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS (OUTER)
NESTED LOOPS
TABLE ACCESS (by index rowid) of table2
INDEX (UNIQUE SCAN) OF PK_TABLE2 (UNIQUE)
INLIST ITERATOR
TABLE ACCESS (BY INDEX ROWID) OF TABLE1
INDEX RANGE SCAN OF PK_TABLE1
VIEW PUSHED PREDICATE OF TABLE3
TABLE ACCESS (BY INDEX ROWID) OF TABLE2
INDEX (UNIQUE SCAN) OF PK_TABLE3
TABLE ACCESS (BY INDEX ROWID) OF TABLE4
INDEX (UNIQUE SCAN) OF PK_TABLE4
INDEX (UNIQUE SCAN) OF PK_TABLE5
VIEW PUSHED PREDICATE OF TABLE3
TABLE ACCESS (BY INDEX ROWID) OF TABLE3
INDEX RANGE SCAN OF PK_TABLE3
The plan for both the unordered and ordered queries are the same!!
Thanks