Make one query from two

Hi,
Could you help optmize this query. If possibel.
I have three tables: GROUP, GROUP_X_USER, USER.
If there is a relation between GROUP and USER, it will be stores at GROUP_X_USER.
Example:
<tt>
GROUP               GROUP_X_USER         USER
1 - GROUP A             1 - 2            1 - MARIO
2 - GROUP B             1 - 3            2 - JOSE
3 - GROUP C             2 - 2            3 - MARIA
</tt>
That means: JOSE can access groups A and B. MARIA, just group A. MARIO does not
have a implicit realation, so he can access all groups (A,B and C).
Clear?
So... I want to create a sql query that return the groups a given user can
access. If in my where clause i use "WHERE USER_ID = 2", it should return 2
lines: GROUP A and GROUP B.
For users that has a impliciti relation at GROUP_X_USER is a peace of cake query.
SELECT GROUP.NAME FROM GROUP, GROUP_X_USER, USER
WHERE GROUP.GROUP_ID = GROUP_X_USER.GROUP_ID
AND GROUP_X_USER.USER_ID = 2;But, this query does not work for user MARIO.
Currently, I´m doing this:
SELECT count(*) INTO wcount FROM GROUP_X_USER WHERE USER_ID = 1;
if wcount = 0 then
    SELECT GROUP.NAME FROM GROUP;
else
    SELECT GROUP.NAME FROM GROUP, GROUP_X_USER, USER
     WHERE GROUP.GROUP_ID = GROUP_X_USER.GROUP_ID
       AND GROUP_X_USER.USER_ID = 2;
end if;
Can I do this with just one query, whithout need do a count and a "if"?
PS: My query is actually more complex than this, has 5 tables and a lot of "conditions".
I made this example just to explain what I need. The goal is to have just one query
to mantein, because, today, if it needs to change, I need to do it two times.
Edited by: mcardia on 30/03/2010 12:15
Edited by: mcardia on 30/03/2010 12:16

Clear?Sorry, but
No,since:
1) you forgot to mention your database version (the result of select * from v$version; )
2) you've posted unformatted code, which you can (and should) easily adjust by adding the tag before and after your examples.
    see: http://forums.oracle.com/forums/help.jspa for more explanation regarding using tags (scroll a bit down)
    we have no clue regarding columns and/or datatypes you're using now.
3) please try to put up a small but concisive testcase, using CREATE TABLE and INSERT INTO statements and the results you want from
    that.
Can I do this with just one query, whithout need do a count and a "if"?
It probably can be done, but you'd need to be more clear and specific first regarding the already mentioned points.

Similar Messages

Make one query of two queries?

I have these two quries, how can I at the best way get the same result in one query?
--Querie 1
SELECT 'm' || Year_Month As Month,
Sum(NET_SALES) As "Net Sales"
FROM tlbInfo
WHERE Year_Month Between '200709' And '200801'
And record_type IN ('A','X')
Group by Year_Month
Order by Year_Month
--Querie 2
SELECT 'm' || Year_Month As Month,
Sum(STANDARD_MANUF_COST)/Sum(Case When SALES_QUANTITY <> '0' Then SALES_QUANTITY End) As "Cost",
to_char(100*((sum(NET_SALES)-sum(STANDARD_MANUF_COST))/sum(NET_SALES)) || '%') As "Margin"
FROM tlbInfo
WHERE Year_Month Between '200709' And '200801'
And record_type NOT IN ('A', 'X')
And Net_Sales >0
Group by Year_Month
Order by Year_Month

why did you change the 1st query? now you're further from getting the correct answer (not to mention that now my previous post makes no sense).
put the 1st query back to the way it was. then change the 2nd query to remove the record_type and net_sales criteria from the where clause, and instead put the logic into the select, via a CASE, like the first query is written.
now look at the two new queries, and ask yourself, what's so hard about making one query?
and don't do this work in the original post. do it in notepad. and if you want to post the new queries here to get further help, hit the REPLY button, NOT THE EDIT BUTTON

Is there any way to work on one Catalog from two computers simultaenously?

Is there any way to work on one Catalog from two computers simultaenously?
I have a catalog with 7000 images we have to process / crop / etc. and I was trying to find a way that two of us could work on the images at the same time.
Thanks!

No, the LR catalog is not multi user capable (and cannot reside on a network share without tricks). What you could do is
export part of the catalog (a subset of all images) into a separate catalog ("Export as Catalog", without exporting the negatives)
work on different images in the two catalogs at the same time
re-import ("Import from Catalog") the exported work after adjustments on it is finished
Beat

I have deleted one query from production server

some one have deleted one query from production server. i would like to know who has deleted that query.
pls let me ASAP.
regards
raja

Hi,
You can get the time stamp( when it was deleted) from table RSRDELREPORTS. Based on that time check with basis guys to get the more info.
regards,
raju

Copy one query from one infocube to another infocube

Hi BW Experts,
Hope all of them are doing good.How can we copy one query from one infocube to another infocube.
I have infocube1 having one query ,I NEED to copy that query to infocube2? How?
Could you please help me out?
Thanks

Use transaction RSZC in BW (BTW you're addressing BW experts in a BO forum better post in the Bex section of the BW forum)

How to make one dvd from several camcoder dvd's

How to make one dvd from several camcoder dvd's. I got a dvd camcoder, to be able to play them into DVD player I had to finalize my disks. Now I got several DVD's whith short files and terrible custom camera menu. I'd like to make it more presentable.
The main problem is I dont want to recode my vob files. So Final Cut and other editing soft is bad idea.
Is there any soft on mac I could reauthor my DVD's?

How to make one dvd from several camcoder dvd's. I got a dvd camcoder, to be able to play them into DVD player I had to finalize my disks. Now I got several DVD's whith short files and terrible custom camera menu. I'd like to make it more presentable.
The main problem is I dont want to recode my vob files. So Final Cut and other editing soft is bad idea.
Is there any soft on mac I could reauthor my DVD's?

Querying from two different databases

Hi All,
I have a Adhoc requirement where in i have to do a select from two tables, but one of them resides on a different database.
The tables have a common primary key, so i can do a one to one match. Is there anyway to query them without using a DB link?

Hi!
I think - if u create a global synonym using that dblink - next time u can use that synonym and u don't have to use any dblink further.
Regards.
Satyaki De.

Combining Data into one Cube from two Data-sources..

Dear Experts,
I am pulling data from two data sources and trying to combine in one Info-Cube. The data are like
Data-Source 01
1. GUID --Common
2.Document No ( User Entry)
3.Dist. Channel
4.Transaction Type
5.Date and Quantity
Data-Source 02
1.GUID -- Common
2.Billing Document ( If User drill down according to Document No , Billing Document should come in the report )
3.Billing date
4.Net Value
Out of the datas , The GUID is common between the 2 data-sources. I was thinking that, tha data will take according to its place and If i select the Document No in Report, it will atomatically fetch all the data like Tran type, dist ch, Billing Document No , Billing date.. .
The problem is , in the report Tha data is not coming as I was thinking.
And Another problem is , In future I need to create a Multiprovider between the above mentioned Info-cube and One ODS. And DOCUMENT NO is common in Cube and ODS.
Please Suggest,
How can I proceed for the following requirement.
Thanks,
Sanjana

Hi Sanjana,
In your case cube will create a problem because it will have multiple records . For example :
Data-Source 01 :
1. GUID -- 101
2.Document No - 999
3.Dist. Channel - DL
4.Transaction Type - GPRO
5.Date and Quantity - 20.02.2011 & 20
Data-Source 02
1.GUID -- 101
2.Billing Document - 6000
3.Billing date - 03.03.2011
4.Net Value - 500
Your cube will have 2 records . And your requirement is to show above two records in 1 record in the report .
Why dont you make an ODS in between , where you can put GUID as the Key field and rest all the fields as data fields. Create 2 transformations to this DSO from the 2 datasources . And let it get updated one by one . Your DSO will have 1 record only . Now either you do reporting on this DSO or take the data to the cube .
Hope the above reply was helpful.
Kind Regards,
Ashutosh Singh
Edited by: Ashutosh Singh on May 19, 2011 1:34 PM

How to Make a report from two cubes

Hi, guys,
i'm trying to create a report use the data from two different cubes, but that's a problem. what i want to do is like this:
From Cube A, i want to get the total machine number for each type of machine:
Type machine | Total machine
Type A 10000
Type B 15000
Type C 40000
And from Cube B, i want to get the machine amount which has an error, also group by machine type:
Type machine | Broken machine amount
Type A 50
Type B 15
Type C 100
So what can i do to create a report like below:
Type machine | Broken machine amount | Broken machine ratio(%)
Type A 50 0,5
Type B 150 1
Type C 100 0.25
Thanks a lot,

Hi, Nicolae, use ur idea, i can get the table u show:
Type | Machine amount | cube
Type A 10000 'cube A'
Type B 15000 'cube A'
Type C 40000 'cube A'
Type A 50 'cube B'
Type B 15 'cube B'
Type C 100 'cube B'
but with this table how can i get the table which i want:
Type | Machine amount | % of machine
Type A 50 0.5
Type B 15 0.1
Type C 100 0.25
I have no idea, 2 things in ur idea have to be considered:
1. just show the record of 'cube B'
2. to get the data that cube B machine amount/ cube A machine amount
Need your detail explain, thanks

How can I make a photo from two merged projects available for a book that I started using one project ? They merge but the book only allows me to access the original project photos even though the merged photos can be seen in the project ?

I have successfully merged two projects in order to add new photos for my book. The library available to the book which has 26 completed pages accesses only the original project library. Has anyone else had this problem ? It is frustrating to see the photos I want in the merged project file, but to have only the original photos when I started the book available in the book library.

That will work, but it is the long way 'round.
As Frank noted, Book Albums are just specialized Albums. Albums can show any Image in your Library, regardless of where the Album is located on the Library Inspector, and regardless of which Project contains the Image. You can put your Albums anywhere (in a Folder called "Books" for example). You can put Images from anywhere in your Library in any Album.
The easy way 'round:
- Make a new Book Album
- Select Images you may want to put in your Book. Flag them.
- Select some more. Flag them.
- Go to the "Flagged" container (listed near the top of the Library Inspector), select all, and drag-and-drop them to your Book Album.
- Repeat as needed, or just drag-and-drop directly into the Book Album.
Note that you can remove any Image from any Album (including Books) by selecting and hitting the "{Delete}" key. This _does not remove the Image from the Project that contains it, or from the Library.
Note, too, that once the Images are in your Book Album, you then put them in the Book you are creating for publication. The Book Album contains the superset of Images with which you populate the Book itself. The Book Album also holds the Book.
A good (imho ) introduction to the parts of Aperture can be found in this short guide I wrote.
The User Manual is helpful. Here is
the chapter on making Books.
From that chapter, here is
the section detailing how to create and populate a Book Album.

Can we make one query to get same results from 3 tables

CREATE TABLE TABLE1 (NODEID VARCHAR2(4));
CREATE TABLE TABLE2 (NODEID VARCHAR2(4));
CREATE TABLE TABLE3 (NODEID VARCHAR2(4));
INSERT INTO TABLE1 VALUE('1004');
INSERT INTO TABLE1 VALUE('1004');
INSERT INTO TABLE1 VALUE('1002');
INSERT INTO TABLE1 VALUE('1002');
INSERT INTO TABLE1 VALUE('1001');
INSERT INTO TABLE1 VALUE('1001');
INSERT INTO TABLE1 VALUE('1006');
INSERT INTO TABLE1 VALUE('1006');
INSERT INTO TABLE1 VALUE('1005');
INSERT INTO TABLE1 VALUE('1005');
INSERT INTO TABLE2 VALUE('1004');
INSERT INTO TABLE2 VALUE('1004');
INSERT INTO TABLE2 VALUE('1004');
INSERT INTO TABLE2 VALUE('1002');
INSERT INTO TABLE2 VALUE('1002');
INSERT INTO TABLE2 VALUE('1002');
INSERT INTO TABLE2 VALUE('1002');
INSERT INTO TABLE3 VALUE('1001');
INSERT INTO TABLE3 VALUE('1001');
INSERT INTO TABLE3 VALUE('1006');
INSERT INTO TABLE3 VALUE('1006');
INSERT INTO TABLE3 VALUE('1005');
INSERT INTO TABLE3 VALUE('1005');
INSERT INTO TABLE3 VALUE('1004');
INSERT INTO TABLE3 VALUE('1004');
INSERT INTO TABLE3 VALUE('1004');
INSERT INTO TABLE3 VALUE('1002');
INSERT INTO TABLE3 VALUE('1002');
INSERT INTO TABLE3 VALUE('1002');
INSERT INTO TABLE3 VALUE('1002');
Select count(*), count(distinct nodeid)
from table1, table2,table3
where table1.nodeid=table2.nodeid and table1.nodeid=table3.nodeid;
Select count(*), count(distinct nodeid)
from table1, table3
where table1.nodeid=table2.nodeid;
Select count(*), count(distinct nodeid)
from table2, table3
where table2.nodeid=table3.nodeid;

Aside from your insert statements not working... (should be as follows)...
DROP TABLE TABLE1;
DROP TABLE TABLE2;
DROP TABLE TABLE3;
CREATE TABLE TABLE1 (NODEID VARCHAR2(4));
CREATE TABLE TABLE2 (NODEID VARCHAR2(4));
CREATE TABLE TABLE3 (NODEID VARCHAR2(4));
INSERT INTO TABLE1 VALUES('1004');
INSERT INTO TABLE1 VALUES('1004');
INSERT INTO TABLE1 VALUES('1002');
INSERT INTO TABLE1 VALUES('1002');
INSERT INTO TABLE1 VALUES('1001');
INSERT INTO TABLE1 VALUES('1001');
INSERT INTO TABLE1 VALUES('1006');
INSERT INTO TABLE1 VALUES('1006');
INSERT INTO TABLE1 VALUES('1005');
INSERT INTO TABLE1 VALUES('1005');
INSERT INTO TABLE2 VALUES('1004');
INSERT INTO TABLE2 VALUES('1004');
INSERT INTO TABLE2 VALUES('1004');
INSERT INTO TABLE2 VALUES('1002');
INSERT INTO TABLE2 VALUES('1002');
INSERT INTO TABLE2 VALUES('1002');
INSERT INTO TABLE2 VALUES('1002');
INSERT INTO TABLE3 VALUES('1001');
INSERT INTO TABLE3 VALUES('1001');
INSERT INTO TABLE3 VALUES('1006');
INSERT INTO TABLE3 VALUES('1006');
INSERT INTO TABLE3 VALUES('1005');
INSERT INTO TABLE3 VALUES('1005');
INSERT INTO TABLE3 VALUES('1004');
INSERT INTO TABLE3 VALUES('1004');
INSERT INTO TABLE3 VALUES('1004');
INSERT INTO TABLE3 VALUES('1002');
INSERT INTO TABLE3 VALUES('1002');
INSERT INTO TABLE3 VALUES('1002');
INSERT INTO TABLE3 VALUES('1002');and you're queries not working...
SQL> Select count(*), count(distinct nodeid)
2 from table1, table2,table3
3 where table1.nodeid=table2.nodeid and table1.nodeid=table3.nodeid;
Select count(*), count(distinct nodeid)
ERROR at line 1:
ORA-00918: column ambiguously definedI'm guessing you want:
SQL> Select count(*), count(distinct table1.nodeid)
2 from table1, table2,table3
3 where table1.nodeid=table2.nodeid and table1.nodeid=table3.nodeid;
COUNT(*) COUNT(DISTINCTTABLE1.NODEID)
50 2You haven't explained to us what database version you are using or what you want the output to look like when these 3 queries are combined.
Please read: {message:id=9360002} and learn to post a good question, and use {noformat}{noformat} tags to show your code/data.

CR 4 Ent, Xcelsius and WebI - one query and two different results

Dear Sirs,
I'm using BO 4.0 platform and 3 tools: Crystal Reports for Enterprise, WebI and Dashboard Design (Xcelsius).
I have one question, because in my opinion in this solution is one inconsistency.
I have one table in SQL Server (like below):
IDRec (autoinc) | Col1 (varchar)   | Col2 (int)
1      A    1
2      A    1
3      B    2
4      B    2
I have one universe with ONLY 3 dimensions (IDRec, Col1, Col2). I haven't any measures.
And:
1. Using CR 4 Ent and I create report using 3 dimensions (IDRec, Col1, Col2).
I make sure, that I selected option: "Retrieve duplicate rows"
If I have 3 columns in details I have 4 records (4 rows) in my report, when I have 2 columns (Col1 Col2) I have only 2 records (2 rows) in report (Page view).
So universe (or something like this) use option DISTICT (I think).
2. When I use WebI (14.0.2) I have the same situation.
3. When I use Dashboard Desigm (Xcelsius 2011) I have 4 rows !!! ALWAYS - it doesn't matter I selected "Retrieve duplicate rows" or not !!!
When I look to "View script" window, I not see DISTINCT clause.
Am I doing something wrong?
I can change the settings so as to be able to see duplicate records?
It's very important for some calculation, especially in Crystal Reports (for Enterprise).

Hi,
Thanks for the response...But the result is deviating from the expected..
       expected is
     if we have first query returns              second query returns
              1                                                     4
              2                                                     5
              3                                                     6
     these two queries result should be in one table , with first query result in first column and second query result in second column.
   The two queries fetching data from same table(category table as in my post) with different search criteria( in where clause).......Regards,
Rakesh.

I am trying to make one query

Hello,
can I make these two queries as one?
           select cv_id,to_char(rDate,'Month dd, yyyy') from jobResponses
           where job_id=28 and responseStatus=0 order by rDate desc
               select
               c.name, nationality, gender, dob, cvName
               from users a, cvProperties b, cvDetails c, countries d
               where a.user_id=b.user_id and b.cv_id=c.cv_id and a.country_id=d.country_id
               and a.userType not like '2' and isEdited=0 order by cvDate desc;My tables:
SQL> desc jobResponses
Name                                      Null?    Type
JR_ID                                     NOT NULL NUMBER(14)
JOB_ID                                             NUMBER(14)
CV_ID                                              NUMBER(14)
RDATE                                              TIMESTAMP(0)
CSDATE                                             TIMESTAMP(0)
RESPONSESTATUS                                     NUMBER(5)
SQL> desc users
Name                                      Null?    Type
USER_ID                                   NOT NULL NUMBER(14)
AGREE_ID                                           NUMBER(14)
COUNTRY_ID                                         NUMBER(14)
REGISTRATIONDATE                                   TIMESTAMP(0)
USERTYPE                                           NUMBER(2)
STATUS                                             NUMBER(2)
NAME                                               VARCHAR2(36)
LOGIN                                              VARCHAR2(110)
PASSWORD                                           VARCHAR2(15)
FOLDERID                                           NUMBER(14)
FOLDERCREATIONDATE                                 DATE
SQL> desc cvProperties
Name                                      Null?    Type
CV_ID                                     NOT NULL NUMBER(14)
USER_ID                                            NUMBER(14)
CAT_ID                                             NUMBER(14)
EMPLOYERORAGENT_ID                                 NUMBER(14)
STAFF_ID                                           NUMBER(14)
CVDATE                                             TIMESTAMP(0)
CVNAME                                             VARCHAR2(230)
STATUS                                             NUMBER(3)
PREV_CVID                                          NUMBER(14)
ISEDITED                                           NUMBER(2)
CVSOURCE                                           NUMBER(2)
SQL> desc cvDetails
Name                                      Null?    Type
CD_ID                                     NOT NULL NUMBER(14)
CV_ID                                              NUMBER(14)
NATIONALITY                                        VARCHAR2(230)
OBJECTIVES                                         VARCHAR2(3900)
NAME                                               VARCHAR2(230)
FATHERNAME                                         VARCHAR2(230)
MOTHERNAME                                         VARCHAR2(230)
DOB                                                VARCHAR2(230)
GENDER                                             VARCHAR2(230)
MSTATUS                                            VARCHAR2(230)
PASSPORTNO                                         VARCHAR2(161)
EMAIL                                              VARCHAR2(230)
PHONE                                              VARCHAR2(75)
MOBILE                                             VARCHAR2(25)
ADDRESS                                            VARCHAR2(2500)
STATE                                              VARCHAR2(230)
ZIPCODE                                            VARCHAR2(230)
CITY                                               VARCHAR2(230)
EDUCATION                                          CLOB
EXPERIENCE                                         CLOB
SKILLS                                             CLOB
LANGUAGES                                          VARCHAR2(1400)
HOBBIES                                            VARCHAR2(3200)
ACHIEVEMENTS                                       VARCHAR2(3900)
REFERENCES                                         VARCHAR2(3900)
SQL> desc countries
Name                                      Null?    Type
COUNTRY_ID                                NOT NULL NUMBER(14)
COUNTRY                                            VARCHAR2(50)
SQL>Thanks in anticipation

Hello,
SQL*Plus: Release 11.2.0.1.0 Production on Sat Mar 31 15:38:27 2012
create table countries(
country_id number(14) primary key,
country varchar2(50)
insert into countries(country_id,country) values(1,'Australia');
insert into countries(country_id,country) values(2,'Newzealand');
create table Users(
user_id number(14) primary key,
country_id number(14) constraint Users_fk1 references countries(country_id),
userType number(2),
status number(2),
name varchar2(36),
login varchar2(110),     /**login will be email**/
password varchar2(15)
insert into users values(1,2,1,1,'Test','[email protected]','testpass');
insert into users values(2,1,1,1,'Test1','[email protected]','testpass1');
insert into users values(3,2,3,1,'employer','[email protected]','testpass1');
create table CvCategories(
cat_id number(14) primary key,
category varchar2(900));
insert into cvCategories values(1,'Doctors');
insert into cvCategories values(2,'Nurses');
create table CvProperties(
cv_id number(14) primary key,
user_id number(14) constraint Cv_fk1 references users(user_id),
cat_id number(14) constraint Cv_fk2 references CvCategories(cat_id),
cvName varchar2(230),
status number(3),
prev_cvId number(14),
isEdited number(2)
insert into cvProperties values(1,2,1,'testCV.doc',1,0,0);
insert into cvProperties values(2,1,2,'testCVs-1.doc',1,0,0);
create table cvDetails(
cd_id number(14) primary key,
cv_id number(14) constraint education_fk references CvProperties(cv_id),
nationality varchar2(230),
name varchar2(230),
dob varchar2(230),
gender varchar2(230)
insert into cvDetails values(1,1,'Test Country','Test Name','31-Feb-2012','Female');
insert into cvDetails values(2,2,'Test Country-1','Test Name-1','31-Feb-2012','Male');
create table jobProperties(
job_id number(14) primary key,
user_id number(14) constraint jobProperties_fk2 references users(user_id),
status number(3),
prev_jobId number(14),
isEdited number(2)
insert into jobProperties values(1,3,1,0,0);
insert into jobProperties values(2,3,1,0,0);
insert into jobProperties values(3,3,1,0,0);
create table JobResponses(
jr_id number(14) primary key,
job_id number(14) constraint JobResponses_fk1 references jobProperties(job_id),
cv_id number(14) constraint JobResponses_fk2 references CvProperties(cv_id),
rDate timestamp(0) default sysdate,
responseStatus number(5)
insert into jobResponses(jr_id,job_id,cv_id,responseStatus) values(1,1,1,0);
insert into jobResponses(jr_id,job_id,cv_id,responseStatus) values(2,1,2,0);
insert into jobResponses(jr_id,job_id,cv_id,responseStatus) values(3,2,1,0);Thanks again

I want to SPLIT A CLIP. That's all. Make one piece into two.

I am looking for the equivalent of the 'Command-T' function in iMovie HD. This is one of the most commonly used actions in processing video, and apparently it doesn't exist. The 'Split Clip' command in iMovie 09 simply does not work for me. At all. EVER. I have yet to find an instance where it isn't grayed out. All I want to do is split a clip, and add a transition between the two.
While I'm on my soap box, the revisions of iMovie since V.6 are without a doubt the most counter-intuitive, joy-sucking applications ever produced by Apple. iMovie made it FUN to edit video; these last two versions make it frustrating and cumbersome.
To reiterate, I want to SPLIT A CLIP INTO TWO PIECES. If anyone here can provide instructions for doing so (as opposed to explaining why it is NOT possible), then kindly do so. Thank you.

Trey, I seriously feel your pain. I'm editing a wedding project in IM '09 and it is literally taking me twice as long to complete the project than it would have in HD. You should be able to make splits in the event browsers, but that's too much like common sense for Apple. They completely mucked up the editing process by doing away with iMovie HD, which by the way is not working well in Snow Leopard.
As for split, I suggest the same thing as above. Put the playhead where you want the split. Us shortcut key, ShiftCommandS. I had problems with it this morning, but it finally kicked in.
I also had a problem when trying delete a selection of a clip. When I clicked delete, it deleted the selection plus the extracted audio of the entire clip from which I took the selection. Not sure what caused that.
I have FCE installed, so maybe I’ll haul it out and try using it for my next project. IM '09 is a pain. I'm not even sure how it's more accessible to general users.

One query on two databases acts differently.

Hi - this is a doozy. I'd Ask Tom but I think he's busy...
Database-A = PRODUCTION
Database-B = TEST
Both databases run 9.2.0.8, the parameter files in use are the same (so all optimizer settings etc are equal). The table spaces were created equally.
One difference is that 'A' uses an overloaded SAN and 'B' is on a different one with plenty of bandwidth - though the problem sounds to me like a database optimizer issue, anyway...
'B' schema was created from export/import of 'A'.
The schemas have both been analyzed the same way (and both do each night).
The problem is here that on 'B' the query uses the PK index on one of the tables whereas on 'A' the same query does not use the PK index.
Note: There are slightly fewer rows in 'B' as 'A' is prod and getting updated.
Already ruled out:
optimizer_index_cost_adj: this is set to 100 on both databases and dropping it to 50, 10, 5 and 1 on 'A' does not make the optimizer favour the index.
Stored outlines may not be an option for this query as we cannot alter the application code and the query changes....
Any ideas where should I look next?
Ta!

Thanks to both the above comments.
I did think about the data clustering but I kept
coming back to why the optimizer on 'A' said "I want
a full table scan" and 'B' says, "oh hey, there's an
index, I'll use that".This has nothing to do with clustering data.
http://download.oracle.com/docs/cd/B19306_01/server.102/b14237/statviews_1069.htm#i1578369 - look for 'Clustering Factor'
Also see Note:39836.1
IN a nutshell - it tells the system how bad an index could be when considering the organization of the table. A table (not index) reorg could easily change the clustering factor, give the same index for the same table with exactly the same data (but different row order) a completely different value.

Make one query from two

Similar Messages

Maybe you are looking for