Union All vs. View Creation Problem

Here is my problem basicly;
I have three table.
select * from table1;
select * from table2;
select * from table3;
I need to create one view by combining those tables with 'union all' word.
create or replace view view1 as
select * from table1;
union all
select * from table2;
union all
select * from table3;
Then, I want to write a query as;
select * from view1 where column1 = x;
When I did this, view1 first combines all the rows of the tables, then searches x value in all the data. That's OK. However I want to first search x value in the tables, and combine the results of each table and get the result rows.
I mean, my query "select * from view1 where column1 = x;" need to act as
select * from table1;
column1 = x
union all
select * from table2;
column1 = x
union all
select * from table3;
column1 = x
How can I do this?
Thanks everyone..

SnigRO
I think what you mean is that your query is sequentially full scanning all three tables, and you would like instead for it to use index scans.
Show the plan for the query. The predicate can be pushed down, but whether it makes any difference may depend:
- statistics on each table (full scan may be faster than index scan)
- are the various column1 columns indexed
- certain documented or hidden optimiser parameters can encourage or discourage push down: eg optimizercost_based_transformation
- Oracle version: there have been known bugs that interfere with push down (eg Bug 3723007 - Predicate push down may not occur in 10g - reported in 10.1)
HTH
Regards Nigel

Similar Messages

  • Probelm with Union all in view

    I'm tring to create a view by using the following:
    SELECT MASTER_IDX.MASTER_IDX_ID, CREATE_BY AS PERSON, CREATE_DATE AS WORKDATE, 'CREATE' AS WORKTYPE
    FROM MASTER_IDX
    WHERE CREATE_BY <> 'IQS'
    UNION ALL
    SELECT MASTER_IDX.MASTER_IDX_ID, INDEX_BY AS PERSON, INDEX_DATE AS WORKDATE, 'INDEX' AS WORKTYPE
    FROM MASTER_IDX
    WHERE INDEX_BY <> 'IQS'
    UNION ALL
    SELECT MASTER_IDX.MASTER_IDX_ID, VER_BY AS PERSON, VER_DATE AS WORKDATE, 'INDEX' AS WORKTYPE
    FROM MASTER_IDX
    WHERE VER_BY <> 'IQS'
    UNION ALL
    SELECT MASTER_IDX.MASTER_IDX_ID, MOD_BY AS PERSON, MOD_DATE AS WORKDATE, 'INDEX' AS WORKTYPE
    FROM MASTER_IDX
    WHERE MOD_BY <> 'IQS'
    UNION ALL
    SELECT MASTER_IDX.MASTER_IDX_ID, MAILBACK_BY AS PERSON, MAILBACK_DATE AS WORKDATE, 'INDEX' AS WORKTYPE
    FROM MASTER_IDX
    WHERE MAILBACK_BY <> 'IQS'
    UNION ALL
    SELECT IDX_IMAGE.MASTER_IDX_ID, SCANNED_BY AS PERSON , SCANNED_DATETIME AS WORKDATE, 'SCANNED' AS WORKTYPE
    FROM IDX_IMAGE
    WHERE SCANNED_BY <> 'IQS'
    when I run this in the sql developer environment I get the results that I expect. When I try to put this into a view I get this error:
    java.lang.NullPointerException
    Can anyone please look at this and give me a clue as to what i'm doing worng?
    Thanks!

    Can you give a bit more detail about which SQL Developer version you are using, what you mean by "put this into a view" (using the "New View ..." window?) and just what you were doing when you got the NPE (ie Test Syntax or OK to create or something else)?
    theFurryOne

  • Performing UNION/UNION ALL in views

    I have two databases (DB1 AND DB2) in same server. I have to create two views (V1 AND V2) respectively in DB1 and DB2. And I have to make an union/union all operation in these views (V1 and V2). Is this possible.???.
    Or can I create a single view instead of two views in such a way that the single view contains an union/union all operation between the data taken from two DBs (DB1 and DB2).?
    Any ideas or suggestions on these questions. Your help is well appreciated.
    Thanks in advance.

    Or can I create a single view instead of two views in such a way that the single view contains an union/union all operation between the data taken from two DBs (DB1 and DB2).?
    Make sure you do very thorough testing if you plan to use UNION since that requires a sort operation. And that means that the data from one query has to be sent to the other server before it can be sorted.
    So a query issued on DB1 might sent the DB1 data to DB2 and then send the sorted result back to DB1. But the same query issued on DB2 would already have the DB2 data so after the DB1 data is sent to DB2 and sorted the query is done.
    That is, in one case the entire result set may be sent over the network from one server to the other instead of just one table.

  • View Creation Problem

    I am facing a problem to creating a view.
    My table data like
    row     col
    1     1
    1     2
    1     3
    1     4     
    1     5
    1     6
    1     7     
    1     8
    1     9
    1     10
    2     1
    2     2
    2     3
    2     4     
    2     5
    2     6
    2     7     
    2     8
    2     9
    2     10
    I want to create a view like this as fields
    row,col1,col2,col3,col4,col5,col6,col7,col8,col9,col10
    problem is sometimes col values upto 10 and some times col values less than 10
    Suppose my data is
    row     col
    1     1
    1     2
    1     3
    1     4     
    1     5
    1     10
    2     1
    2     2
    2     3
    2     4     
    2     5
    then also i want view like
    row,col1,col2,col3,col4,col5,col6,col7,col8,col9,col10
    and col6,col7,col8,col9,col10 are as null values.
    If any body have idea to solve this issue, please help me
    Advance Thanking,
    siju Abraham

    If you search on this forum for the word "pivot", you'll see lots of examples of queries using max(decode(...)).
    Regards,
    Rob.

  • Inventory on Hand Report - Using Union ALL - need two fields to be part of the group

    Hi all,
    I have created an Inventory "on Hand" Report that takes the Current Inventory from the Item Entry table and the Sales from Unposted Sales Line table and Transfers in and out from the Unposted Transfer Line table.  I have joined the tables
    using the UNION ALL function. 
    My problem is that the Transfer table has two locations whereas the other tables only have one.  I am grouping on Location code from the Item Entry table which is equivalent to BOTH Transfer from location and the Transfer to.    
    As an example, there are 15lbs of inventory for Product A in Location #1 with a transfer out of 15 lbs.  The Transfer out is going to Location #2
    I don't know how to write the query or set up the group so that it recognizes both the Transfer to and the Transfer From fields
    I want the report to look similar to the one below but I can only get it to show one of the locations.  Is there some way to use the Union function and have one field in the first table be or equivalent to two fields in another?
    Location   Code
    Item No.
    Lbs
    Sales Orders
    Transfer Out
    Transfer In
    Available   Inventory
    Location #1
    Product A
    15
    -15
    Location #1
    15
    -15
    Location #2
    Product A
    15
    15
    Location #2
    15
    15

    Hi Igor,
    You can get a custom sort order added to your IP column without the need for a second column.
    Consider that the sorting is done strictly left-to-right across a string in the column. The string can be any valid HTML content. So, you could wrap your string within, say, a SPAN tag and add an attribute to that tag that contains the sort order you need before the text that is displayed to the user. As long as the attribute is correctly structured (that is, all instances are of the same length, for example), then sorting will work correctly. For example:
    SELECT
    '<span title="' || PAD_IP_ADDRESS(IP) || '">' || IP || '</span>' Y
    FROM ...Now you need to ensure that the PAD_IP_ADDRESS() function returns the correct values. In IP addresses, you have anything from "0.0.0.0" to "255.255.255.255". To get them to sort "numerically", you need to pad one or two digit numbers to get three digit numbers for each value - so, "0.0.0.0" becomes "000.000.000.000". You could create a function to do this - something like:
    CREATE OR REPLACE FUNCTION PAD_IP_ADDRESS
      pIP IN VARCHAR2
    RETURN VARCHAR2
    IS
      vIP VARCHAR2(15);
      vTEMP APEX_APPLICATION_GLOBAL.VC_ARR2;
      vSEP VARCHAR2(1);
    BEGIN
      vSEP := '';
      vIP := '';
      vTEMP := APEX_UTIL.STRING_TO_TABLE(pIP,'.');
      FOR x IN 1..vTEMP.COUNT
      LOOP
        vIP := vIP || vSEP || TRIM(TO_CHAR(TO_NUMBER(vTEMP(x)),'000'));
        vSEP := '.';
      END LOOP;
      RETURN vIP;
    END;The output from this would look something like:
    &lt;span title="001.001.001.001"&gt;1.1.1.1&lt;/span&gt;
    &lt;span title="002.255.255.255"&gt;2.255.255.255&lt;/span&gt;
    &lt;span title="010.001.199.098"&gt;10.1.199.098&lt;/span&gt;Andy

  • Materialized view issue with spatial index and UNION all.

    Hi guys,
    I'm trying to build the following materialized view:
    create materialized VIEW MV_ElectricalStuffs
      refresh fast
      AS
      SELECT jb.ROWID,
        jb.FID,
        JB.NAME_NUMBER
      FROM EL_BUS_BAR jb
      UNION ALL
      SELECT INC.ROWID,
        INC.FID,
        NULL,
        INC.NAME_NUMBER
      FROM EL_INTERNAL_CELL INC;
    I have this error showing up:
    ORA-12015: cannot create a fast refresh materialized view from a complex query
    This is because the table EL_INTERNAL_CELL has a SDO_GEOMETRY column that has a spatial index, whose ddl is
    CREATE INDEX EL_INTERNAL_CELL_S ON EL_INTERNAL_CELL (GEOM) INDEXTYPE IS MDSYS.SPATIAL_INDEX
    PARAMETERS('SDO_INDX_DIMS=2 TABLESPACE=USERS LAYER_GTYPE=COLLECTION');
    When I remove the spatial index from EL_INTERNAL_CELL column GEOM, Oracle is very happy and creates the view.
    Is there a mean however to keep the spatial index in the materialized view?

    I've managed to drop the spatial index prior to create the materialized view and it is ok. After the materialized view creation, I've recreated the spatial index on the table and all ran smooth. Hope nothing will go bad in the future because of this trick on spatial index..

  • Strange behaviour of view based on several tables join with union all

    Dear fellows we r facing a strange problem we have a view based on several tables joined by union all ,when we issue an ordered query on date,rows returned are unusually different than they should be .
    Is oracle has some special behaviour for view based on union all ?
    I m using oracle 8.1.6 on windows 2000
    Kashif Sohail

    Did you ever solve this problem? I have two select statements based on 4 tables and 3 views using about 5 more tables. When I execute each select individually I get exactly what is expected. When I UNION ALL both selects together, I get "no rows returned", however when I UNION them I get exactly what is expected. I should get the same answer for both UNION ALL and UNION. The two select statements are identical, except for one column where I changed the constant to be different. Any thoughts?

  • Oracle doc inconsistent on materialize view with union all and self joins

    First of all, I can't seem to create a materialized view containing self-joins AND union all. Is it possible?
    I checked Oracle 9i (my version: PL/SQL Release 9.2.0.4.0 - Production) documentation and I get different answers (or so it seems to me).
    First I saw this: "The COMPATIBILITY parameter must be set to 9.0 if the materialized aggregate view has inline views, outer joins, self joins or grouping sets and FAST REFRESH is specified during creation..."
    Did you see the part about 'self joins' in there? I did and I was pumped because that seems to say that you CAN have 'self joins' (and my compatibility is 9.2...)
    BUT
    In the very same document I also found "Oracle does not allow self-joins in materialized join views." (rage)
    You can see the document I am speaking of here: http://download-west.oracle.com/docs/cd/B10501_01/server.920/a96520/mv.htm#574889
    Whenever I try to create the mview I get the following error. (
    In any caseORA-01446 cannot select ROWID from view with DISTINCT, GROUP BY, etc.

    First of all, I can't seem to create a materialized view containing self-joins AND union all. Is it possible?
    I checked Oracle 9i (my version: PL/SQL Release 9.2.0.4.0 - Production) documentation and I get different answers (or so it seems to me).
    First I saw this: "The COMPATIBILITY parameter must be set to 9.0 if the materialized aggregate view has inline views, outer joins, self joins or grouping sets and FAST REFRESH is specified during creation..."
    Did you see the part about 'self joins' in there? I did and I was pumped because that seems to say that you CAN have 'self joins' (and my compatibility is 9.2...)
    BUT
    In the very same document I also found "Oracle does not allow self-joins in materialized join views." (rage)
    You can see the document I am speaking of here: http://download-west.oracle.com/docs/cd/B10501_01/server.920/a96520/mv.htm#574889
    Whenever I try to create the mview I get the following error. (
    In any caseORA-01446 cannot select ROWID from view with DISTINCT, GROUP BY, etc.

  • Transaction I view for WebGui(Html)  creation problem

    Hi All,
    I am creating a transaction Iview for WebGui, I am done with I view creation part but when I open the I view its giving 
    <b>No switch made to HTTPS; so it is unsafe to send your password</b>.
    SSO between was and ecc is working fine. I read postings related this issue. No one is able to get the proper solution. I mean why it is happening and how to fix it. Please let me know if any knows how to fix this.
    Thanks
    Keerthi

    Hi keerthi,
    SSO is between???
    and also one more thing check the preview of IVIEW, if iview is perfect then i dont think so there is any problem.
    Thanks
    Purav

  • Using timestamp in union all view

    I am having a silly problem with one of my views. It is a union all with two select statements. There is a timestamp(6) in one select but not in the other so I have a to_timestamp(null) to match up the columns.
    The problem is that no matter how I code the to_timestamp(null,fmt) the structure of the view is always timestamp(0). I even changed the 2nd select so that it uses a timestamp(6) column. If I create 2 views from both selects they have the correct syntax but as soon as I use a union the timestamp loses it precision.
    When I acutally do a select from the view I can still see the milliseconds precision. My real problem occurs in that I have a ref cursor that is of type view_row%rowtype. When I select into the cursor I lose the precision.
    Any ideas?
    thx

    Use the cast function to cast the timestamp to timestamp(6)
    SQL> create table t
    2 ( id number
    3 , ts timestamp(6)
    4 );
    Table created.
    SQL> create view v
    2 as
    3 select id, ts
    4 from t
    5 union all
    6 select 0, cast(to_timestamp(null) as timestamp(6))
    7 from dual;
    View created.
    SQL> desc v
    Name Null? Type
    ID NUMBER
    TS TIMESTAMP(6)
    SQL> insert into t
    2 values (1,systimestamp);
    1 row created.
    SQL> select * from v;
    ID
    TS
    1
    31-AUG-04 02.28.38.342412 PM
    0

  • Materialized view - REFRESH FAST ON COMMIT - UNION ALL

    In a materialized view which uses REFRESH FAST ON COMMIT to update the data, how many UNION ALL can be used.
    I am asking this question because my materialized view works fine when there are only 2 SELECT statement (1 UNION ALL).
    Thank you.

    In a materialized view which uses REFRESH FAST ON COMMIT to update the data, how many UNION ALL can be used.As far as I remember you can have 64K UNIONized selects.
    I am asking this question because my materialized view works fine when there are only 2 SELECT statement (1 UNION ALL).Post SQL that does not work.

  • OAF Export button fetching data in one column - view object using union all

    Dear All,
    Export button showing data in one column from  view object,
    View object is based on mulitple queries with union all ,
    Please let me know the solution for this issue.
    Thanks
    Maheswara Raju

    Maheswara Raju,
    As per my understanding you are not able to export all the View Attribute using export Button. Only the attribute which is used with the item/region will get exported.
    There are few work around in case if you want to export the column without showing on OAF Page. Let me know.
    Cheers
    Gyan

  • Updatable Materialized View with Union ALL

    (please don't ask about db structure)
    DB: 11gR2
    create table table_1  (
        id number primary key,
        val varchar2(100)
    create table table_2  (
        id number primary key,
        val varchar2(100)
    insert into table_1(id) values (0);
    insert into table_1(id) values (2);
    insert into table_1(id) values (3);
    insert into table_1(id) values (4);
    insert into table_1(id) values (5);
    insert into table_2(id) values (10);
    insert into table_2(id) values (12);
    insert into table_2(id) values (13);
    insert into table_2(id) values (14);
    insert into table_2(id) values (15);
    update table_1 set val='Table1 val:'||id;
    update table_2 set val='Table2 val:'||id;
    create view v_table_all as
    select * from table_1
    view V_TABLE_ALL created.
    select * from v_table_all;
    ID                     VAL                                                                                                 
    0                      Table1 val:0                                                                                        
    2                      Table1 val:2                                                                                        
    3                      Table1 val:3                                                                                        
    4                      Table1 val:4                                                                                        
    5                      Table1 val:5                                                                                        
    select column_name, updatable, insertable, deletable
    from user_updatable_columns
    where table_name = 'V_TABLE_ALL'
    COLUMN_NAME                    UPDATABLE INSERTABLE DELETABLE
    ID                             YES       YES        YES      
    VAL                            YES       YES        YES      
    update v_table_all set val='XXX changed' where id = 3;
    1 row updated.
    select * from table_1;
    ID                     VAL                                                                                                 
    0                      Table1 val:0                                                                                        
    2                      Table1 val:2                                                                                        
    3                      XXX changed                                                                                         
    4                      Table1 val:4                                                                                        
    5                      Table1 val:5                                                                                        
    rollback;
    select * from table_1;
    ID                     VAL                                                                                                 
    0                      Table1 val:0                                                                                        
    2                      Table1 val:2                                                                                        
    3                      Table1 val:3                                                                                        
    4                      Table1 val:4                                                                                        
    5                      Table1 val:5                                                                                        
    create or replace view v_table_all as
    select * from table_1
    union select * from table_2;
    view V_TABLE_ALL created.
    select * from v_table_all;
    ID                     VAL                                                                                                 
    0                      Table1 val:0                                                                                        
    2                      Table1 val:2                                                                                        
    3                      Table1 val:3                                                                                        
    4                      Table1 val:4                                                                                        
    5                      Table1 val:5                                                                                        
    10                     Table2 val:10                                                                                       
    12                     Table2 val:12                                                                                       
    13                     Table2 val:13                                                                                       
    14                     Table2 val:14                                                                                       
    15                     Table2 val:15  
    select column_name, updatable, insertable, deletable
    from user_updatable_columns
    where table_name = 'V_TABLE_ALL'
    COLUMN_NAME                    UPDATABLE INSERTABLE DELETABLE
    ID                             NO        NO         NO       
    VAL                            NO        NO         NO       
    trying update:
    update v_table_all set val='XXX changed' where id = 3;
    SQL-Fehler: ORA-01732: Datenmanipulationsoperation auf dieser View nicht zulässig
    01732. 00000 -  "data manipulation operation not legal on this view"
    *Cause:   
    *Action:
    drop view v_table_all;
    view V_TABLE_ALL dropped.all is ok before this point.
    now we want create a new materialized view with some query
    create  materialized view v_table_all
    as
    select * from table_1
    union all select * from table_2 ;
    materialized view V_TABLE_ALL created.
    select column_name, updatable, insertable, deletable
    from user_updatable_columns
    where table_name = 'V_TABLE_ALL'
    COLUMN_NAME                    UPDATABLE INSERTABLE DELETABLE
    ID                             YES       YES        YES      
    VAL                            YES       YES        YES       it seems to be ok with update.
    but...
    update v_table_all set val='XXX changed' where id = 3;
    SQL-Fehler: ORA-01732: Datenmanipulationsoperation auf dieser View nicht zulässig
    01732. 00000 -  "data manipulation operation not legal on this view"
    *Cause:   
    *Action:How can solve this issue??
    Any suggestion

    Looks like user_updatable_columns sort of thinks the MV is just a table - I don't know about that...
    An MV on a single table can be updated - I tried that and it works:
    create materialized view mv_table_1 for update
    as
    select * from table_1;I noticed [url http://download.oracle.com/docs/cd/E11882_01/server.112/e16579/advmv.htm#sthref294]examples stating the UNION ALL needs a "marker" so Oracle can know from the data which source table a row in the MV originates from - like this:
    create materialized view v_table_all for update
    as
    select 'T1' tab_id, table_1.* from table_1
    union all
    select 'T2' tab_id, table_2.* from table_2 ;But that also fails (the "marker" requirement was specifically for FAST REFRESH, so it was just a long shot ;-) )
    What are you planning to do?
    <li>Create the MV.
    <li>Update records in the MV - which then is no longer consistent with the source data.
    <li>Schedule a complete refresh once in a while - thereby overwriting/losing the updates in the MV.
    If that is the case, I suggest using a true table rather than an MV.
    <li>Create table t_table_all as select ... .
    <li>Update records in the table - which then is no longer consistent with the source data.
    <li>Schedule a job to delete table and insert into table select ... once in a while - thereby overwriting/losing the updates in the table.
    In other words a kind of "do it yourself MV".
    I cannot see another way at the moment? But perhaps try in the data warehousing forum - the people there may have greater experience with MV's ;-)

  • Materialized Views Union ALL

    Sample SQL Query for Creating View
    CREATE MATERIALIZED VIEW TRIAL2
    PARALLEL 4
    BUILD IMMEDIATE
    REFRESH COMPLETE
    ENABLE QUERY REWRITE
    AS */
    select maa.INVENTORY_ITEM_ID,maa.ORGANIZATION_ID,maa.SR_INSTANCE_ID from msc_atp_rules ma,
              msc_atp_assignments maa
    where      maa.assignment_type = 3
    AND          maa.ATP_RULE_ID = ma.RULE_ID
    UNION ALL
    select maa.INVENTORY_ITEM_ID,maa.ORGANIZATION_ID,maa.SR_INSTANCE_ID from msc_atp_rules ma,
              msc_atp_assignments maa
    where      maa.assignment_type =3
    AND          maa.ATP_RULE_ID = ma.RULE_ID
    Test Syntax
    The SQL syntax is valid, however the query is invalid or uses functionality that is not supported.
    Declarative query support does not currently include UNION, INTERSECT or MINUS
    As a workaround, I created this view in the database and imported it onto jdev. This process worked totally fine. However, the Test Syntax still gives the same error, which means no modifications can be made on Jdev.
    Is my conclusion right, or am I missing some procedures that might get the UNIONs working?
    Thanks
    Rajiv

    Hi John,
    Thanks.
    How shall we create xdf then for such MV? Instead of xdf can we have sql script shipped.... or is there any other way of shipping such materialized view... In R12 and 11i (after 11.5.10) we used to ship corresponding xdf but in before 11.5.10 we used to ship sql scripts.

  • Inconsistent SQL results when using View with UNION-ALL and table function

    Can any of you please execute the below scripts and check the output. In the table type variable, I am adding 4 distinct object ids, where as in the result, I get only the row pertaining to last id in the table type variable. Same row is returned 4 times (4= number of values in the table type).
    This scenario is occurring in our product with a SQL with exactly same pattern. I could simulate the same issue with the sample script I have provided.
    Database version: 11.2.0.3 Enterprise Edition, Single node
    Thank you.
    CREATE TABLE TEMP_T1 AS SELECT * FROM ALL_OBJECTS;
    CREATE TABLE TEMP_T2 AS SELECT * FROM ALL_OBJECTS;
    UPDATE TEMP_T2 SET OBJECT_ID = OBJECT_ID * 37;
    CREATE UNIQUE INDEX TEMP_T1_U1 ON TEMP_T1(OBJECT_ID);
    CREATE UNIQUE INDEX TEMP_T2_U1 ON TEMP_T2(OBJECT_ID);
    CREATE OR REPLACE VIEW TEMP_T1T2_V AS
    SELECT * FROM TEMP_T1 UNION ALL SELECT * FROM TEMP_T2;
    CREATE OR REPLACE TYPE TEMP_OBJ_TYPE AS OBJECT (OBJ_ID NUMBER);
    CREATE OR REPLACE TYPE TEMP_OBJ_TAB_TYPE IS TABLE OF TEMP_OBJ_TYPE;
    SET SERVEROUTPUT ON;
    DECLARE
    TYPE TEMP_T1T2_V_ROW_TAB_TYPE IS TABLE OF TEMP_T1T2_V%ROWTYPE;
    TEMP_T1T2_V_ROW_TAB TEMP_T1T2_V_ROW_TAB_TYPE;
    TEMP_OBJ_TAB TEMP_OBJ_TAB_TYPE := TEMP_OBJ_TAB_TYPE();
    PROCEDURE ADD_TO_TEMP_OBJ_TAB(OBJ_ID IN NUMBER) IS
    BEGIN
    TEMP_OBJ_TAB.EXTEND;
    TEMP_OBJ_TAB(TEMP_OBJ_TAB.LAST) := TEMP_OBJ_TYPE(OBJ_ID);
    END;
    BEGIN
    ADD_TO_TEMP_OBJ_TAB(100);
    ADD_TO_TEMP_OBJ_TAB(116);
    ADD_TO_TEMP_OBJ_TAB(279);
    ADD_TO_TEMP_OBJ_TAB(364);
    DBMS_OUTPUT.PUT_LINE('=====================');
    FOR I IN TEMP_OBJ_TAB.FIRST..TEMP_OBJ_TAB.LAST
    LOOP
    DBMS_OUTPUT.PUT_LINE('OBJ_ID = '||TEMP_OBJ_TAB(I).OBJ_ID);
    END LOOP;
    DBMS_OUTPUT.PUT_LINE('---------------------');
    SELECT * BULK COLLECT INTO TEMP_T1T2_V_ROW_TAB
    FROM TEMP_T1T2_V VW
    WHERE ((VW.OBJECT_ID) IN (SELECT OBJ_ID
    FROM TABLE(CAST(TEMP_OBJ_TAB AS TEMP_OBJ_TAB_TYPE))));
    FOR I IN TEMP_OBJ_TAB.FIRST..TEMP_OBJ_TAB.LAST
    LOOP
    DBMS_OUTPUT.PUT_LINE('OBJ_ID = '||TEMP_OBJ_TAB(I).OBJ_ID);
    END LOOP;
    DBMS_OUTPUT.PUT_LINE('---------------------');
    IF TEMP_T1T2_V_ROW_TAB.COUNT > 0 THEN
    FOR I IN TEMP_T1T2_V_ROW_TAB.FIRST..TEMP_T1T2_V_ROW_TAB.LAST
    LOOP
    DBMS_OUTPUT.PUT_LINE(TEMP_T1T2_V_ROW_TAB(I).OBJECT_ID||' : '||TEMP_T1T2_V_ROW_TAB(I).OBJECT_NAME);
    END LOOP;
    ELSE
    DBMS_OUTPUT.PUT_LINE('NO ROWS RETURNED!');
    END IF;
    DBMS_OUTPUT.PUT_LINE('---------------------');
    END;
    /

    I can reproduce it:
    SQL*Plus: Release 11.2.0.3.0 Production on Tue Oct 30 14:05:39 2012
    Copyright (c) 1982, 2011, Oracle.  All rights reserved.
    Enter user-name: scott
    Enter password:
    Connected to:
    Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production
    With the Partitioning, OLAP, Data Mining and Real Application Testing options
    SQL> select  *
      2    from  v$version
      3  /
    BANNER
    Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production
    PL/SQL Release 11.2.0.3.0 - Production
    CORE    11.2.0.3.0      Production
    TNS for 64-bit Windows: Version 11.2.0.3.0 - Production
    NLSRTL Version 11.2.0.3.0 - Production
    SQL> CREATE TABLE TEMP_T1 AS SELECT * FROM ALL_OBJECTS;
    Table created.
    SQL>
    SQL> CREATE TABLE TEMP_T2 AS SELECT * FROM ALL_OBJECTS;
    Table created.
    SQL>
    SQL> UPDATE TEMP_T2 SET OBJECT_ID = OBJECT_ID * 37;
    72883 rows updated.
    SQL>
    SQL> CREATE UNIQUE INDEX TEMP_T1_U1 ON TEMP_T1(OBJECT_ID);
    Index created.
    SQL>
    SQL> CREATE UNIQUE INDEX TEMP_T2_U1 ON TEMP_T2(OBJECT_ID);
    Index created.
    SQL>
    SQL> CREATE OR REPLACE VIEW TEMP_T1T2_V AS
      2  SELECT * FROM TEMP_T1 UNION ALL SELECT * FROM TEMP_T2;
    View created.
    SQL>
    SQL> CREATE OR REPLACE TYPE TEMP_OBJ_TYPE AS OBJECT (OBJ_ID NUMBER)
      2  /
    Type created.
    SQL> CREATE OR REPLACE TYPE TEMP_OBJ_TAB_TYPE IS TABLE OF TEMP_OBJ_TYPE
      2  /
    Type created.
    SQL> SET SERVEROUTPUT ON;
    SQL>
    SQL> DECLARE
      2  TYPE TEMP_T1T2_V_ROW_TAB_TYPE IS TABLE OF TEMP_T1T2_V%ROWTYPE;
      3  TEMP_T1T2_V_ROW_TAB TEMP_T1T2_V_ROW_TAB_TYPE;
      4  TEMP_OBJ_TAB TEMP_OBJ_TAB_TYPE := TEMP_OBJ_TAB_TYPE();
      5  PROCEDURE ADD_TO_TEMP_OBJ_TAB(OBJ_ID IN NUMBER) IS
      6  BEGIN
      7  TEMP_OBJ_TAB.EXTEND;
      8  TEMP_OBJ_TAB(TEMP_OBJ_TAB.LAST) := TEMP_OBJ_TYPE(OBJ_ID);
      9  END;
    10  BEGIN
    11  ADD_TO_TEMP_OBJ_TAB(100);
    12  ADD_TO_TEMP_OBJ_TAB(116);
    13  ADD_TO_TEMP_OBJ_TAB(279);
    14  ADD_TO_TEMP_OBJ_TAB(364);
    15  DBMS_OUTPUT.PUT_LINE('=====================');
    16  FOR I IN TEMP_OBJ_TAB.FIRST..TEMP_OBJ_TAB.LAST
    17  LOOP
    18  DBMS_OUTPUT.PUT_LINE('OBJ_ID = '||TEMP_OBJ_TAB(I).OBJ_ID);
    19  END LOOP;
    20  DBMS_OUTPUT.PUT_LINE('---------------------');
    21  SELECT * BULK COLLECT INTO TEMP_T1T2_V_ROW_TAB
    22  FROM TEMP_T1T2_V VW
    23  WHERE ((VW.OBJECT_ID) IN (SELECT OBJ_ID
    24  FROM TABLE(CAST(TEMP_OBJ_TAB AS TEMP_OBJ_TAB_TYPE))));
    25  FOR I IN TEMP_OBJ_TAB.FIRST..TEMP_OBJ_TAB.LAST
    26  LOOP
    27  DBMS_OUTPUT.PUT_LINE('OBJ_ID = '||TEMP_OBJ_TAB(I).OBJ_ID);
    28  END LOOP;
    29  DBMS_OUTPUT.PUT_LINE('---------------------');
    30  IF TEMP_T1T2_V_ROW_TAB.COUNT > 0 THEN
    31  FOR I IN TEMP_T1T2_V_ROW_TAB.FIRST..TEMP_T1T2_V_ROW_TAB.LAST
    32  LOOP
    33  DBMS_OUTPUT.PUT_LINE(TEMP_T1T2_V_ROW_TAB(I).OBJECT_ID||' : '||TEMP_T1T2_V_ROW_TAB(I).OBJECT_NAME);
    34  END LOOP;
    35  ELSE
    36  DBMS_OUTPUT.PUT_LINE('NO ROWS RETURNED!');
    37  END IF;
    38  DBMS_OUTPUT.PUT_LINE('---------------------');
    39  END;
    40  /
    =====================
    OBJ_ID = 100
    OBJ_ID = 116
    OBJ_ID = 279
    OBJ_ID = 364
    OBJ_ID = 100
    OBJ_ID = 116
    OBJ_ID = 279
    OBJ_ID = 364
    364 : I_AUDIT
    364 : I_AUDIT
    364 : I_AUDIT
    364 : I_AUDIT
    PL/SQL procedure successfully completed.
    SQL> column object_name format a30
    SQL> select  object_id,
      2          object_name
      3    from  dba_objects
      4    where object_id in (100,116,279,364)
      5  /
    OBJECT_ID OBJECT_NAME
           100 ORA$BASE
           116 DUAL
           279 MAP_OBJECT
           364 I_AUDIT
    SQL>  Works fine in:
    =====================
    OBJ_ID = 100
    OBJ_ID = 116
    OBJ_ID = 279
    OBJ_ID = 364
    OBJ_ID = 100
    OBJ_ID = 116
    OBJ_ID = 279
    OBJ_ID = 364
    100 : ORA$BASE
    116 : DUAL
    364 : SYSTEM_PRIVILEGE_MAP
    279 : MAP_OBJECT
    PL/SQL procedure successfully completed.
    SQL> select  object_id,
      2          object_name
      3    from  dba_objects
      4    where object_id in (100,116,279,364)
      5  /
    OBJECT_ID OBJECT_NAME
          100 ORA$BASE
          116 DUAL
          364 SYSTEM_PRIVILEGE_MAP
          279 MAP_OBJECT
    SQL> select  *
      2    from  v$version
      3  /
    BANNER
    Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - Production
    PL/SQL Release 11.2.0.1.0 - Production
    CORE    11.2.0.1.0      Production
    TNS for 32-bit Windows: Version 11.2.0.1.0 - Production
    NLSRTL Version 11.2.0.1.0 - Production
    SQL>SY.
    Edited by: Solomon Yakobson on Oct 30, 2012 2:14 PM

Maybe you are looking for