LEADING or ORDERED hint in OWB 10gR2

Hi,
How does one add a LEADING or ORDERED hint to a mapping in OWB 10g R2?
The usual way of adding a hint is to "Configure" the map and add the extraction or loading hint under the table (data source) operators. However, the hints in such a case are specific to table under consideration. A NO_INDEX or a PARALLEL hint can be applied under it.
However, a LEADING or a ORDERED does not necessarily depend on the table but is common to the entire SQL. Although OWB allows attaching the LEADING or ORDERED hint under Map Configure > Table Operators > [Table Name] > Hints > Extraction Hints, it does not attach the correct table aliases.
Has anyone encountered this problem before? If you've found a workaround, please share it.
Thanks, JD

Hi,
In OWB 10.2 it's often not easy to set hints. Often it's only possible to set them correctly using workarounds :-(
The most important is not where you define the hint in OWB but where OWB puts it in the generated code.
So if you want to have the LEADIND or ORDERING hint in your code you could define it at the level of the target table under:
Map Configure > Table Operators > [Target Table Name] > Hints > Loading Hints
Another way to do it would be to generate a dummy filter (1=1) in the mapping and then to define the hint for this filter.
For hints which are not related to a specific table I mostly used the trial-and-error method to find out where I have to define it in OWB.
Good luck

Similar Messages

OWB 10gR2 - some enhancement requests

Hi, I've been working with OWB 10gR2 for approximately a week now, and I've come across a few (minor) bugs, as well as a list of enhancements that I think would make the tool better. I'd like to surface these in this forum so they can be either disagreed with or have workarounds suggested, as well as getting this list in front of Oracle. Also I'm new to the tool, so maybe there are ways to accomplish what I want.
p.s. as a side note - please don't take these as criticisms of the tool. I'm actually very impressed with OWB 10gR2, but just want to surface some ideas that I think would make the tool better.
Thanks,
Scott
---Bugs:
1.     If I enter a quote character in one of the business description fields (i.e. for a cube or dimension), it causes the DDL scripts to blow up. It probably mentions the list of valid characters in the users guide somewhere - but if a character isn't legal, it should be flagged when I try to enter it originally not when deploying
2.     I can enter illegal strings in an expression operator that fail validation there, but the whole map still passes validation. For instance, I put in an expression (for a varchar2 field) using a single quote at the start, but mistakenly with a double quote at the end. I didnt bother to validate this within the expression. The whole map validated fine, but (of course) blew up with the illegal syntax when generating the code.
3.     I get the following error when trying to browse a cube (ROLAP, not MOLAP) through OWB, but I'm able to pull the cube up fine using either the excel add-in or biBeans: CubeDV_OLAPSchemaConnectionException_ENT_06952???
---Enhancement Requests:
1.     Add a deploy button in control center next to the to default actions and reset actions buttons
2.     Allow an option to have dimensions and cube defaults set to either deploy all, deploy data only, deploy catalog only, deploy aggregations (for cubes). Also, why is the default set to deploy data only? Does deploying the CMW2 metadata involve huge performance impacts or something? It seems like the CMW2 is almost free so it should be turned on by default. Its a pain to use the dimension wizard, but then have to go in to each dimension afterwards to set this option.
3.     When using configure on dims and cubes in control center, the first time the configuration properties dialog pops up, everything is fine. However, when you try to configure the 2nd object, the background focus switches from control center back to design center. This makes navigation very slow if you have lots of items to change. I've noticed this behavior in a few other places as well.
4.     Some dialogs do not have maximize buttons (i.e. dimension wizard dialogs), and default to a very small size. It is possible to have size of dialogs dynamically scaled to screen resolutions, and attempt to size columns so they fit properly. Or better yet, have each dialog remember how large it was last time, and automatically open up to that size?
5.     When generating the autobind relational table while creating a dimension, it seems to be randomizing the order of the attributes within a level (although it does sort the top level attributes first, followed by lower level attributes), at least when using the dimension wizard. It would be nice if the attributes in the relational table came out sorted in the same order as they are specified in the attributes tab.
6.     When using a unmapped display set there doesnt seem to be a way to refresh the set (i.e. remove the items that have since been mapped) unless you select all and then unmapped again. Would be nice if you could simply choose unmapped again and have it refresh the object
7.     While doing a mapping, right click on canvas allows option to generate. Would be nice to also add the validate option to this menu
8.     Validating mappings takes a decent amount of time would be nice if a busy icon was displayed to let user know that OWB hasnt locked up.
9.     default size of job details dialog is too small, and even if I maximize I still have to go through and manually adjust the size of each pane. Similar to enhancement #4
10.     Horizontal size of operators in mapping seems to be determined by the length of the operator name, instead of tying directly to the icon. Looks strange. Very minor
11.     When displaying dimension objects in a mapping:
a.     For the surrogate key columns add an icon to indicate this attribute is a surrogate key. Likewise, since its not legal to map anything to this column, remove the little arrow next to it (the one that turns gray when something is mapped to a column), or perhaps don't even show this column at all.
b.     For the business key columns add some type of icon to show which attributes represent the business natural key (or make color different, etc.)
12.     Similar to #11, for the cube operator in a mapping:
a.     For the surrogate key columns if they cant be mapped to, then dont show them
b.     For the business key columns add some type of icon to show which attributes represent the business natural key
c.     Separate the measure columns from the dimension columns
13.     For the Aggregation operator can you move the <None> default operator to the top of the list. At the bottom, it means you always have to scroll to use the most common operations (which are at the top of the list), i.e. sum and avg, etc.
14.     In the log on prompt, if I choose SQL*Net connection, every time I start it up it defaults to the details being shown (even when I chose no detail last time), with cursor in the host box. I should be able to have the details hidden, and this dialog should ALWAYS default to having the active cursor in the password box on startup
15.     When printing in a mapping, the page setup print dialog is very slow. Also, the print setup dialog shows margins, but you cant adjust them. It would be nice if this could be changed here
16.     Add a print preview icon to all of the toolbars. Printer icon is there, print preview should also be.
17.     In the dimension attributes, we need a way to tag an attribute as being the insert record date or the update record date. The corresponding MERGE SQL should be updated to use these attributes.
18.     It would be nice if the mapping canvas defaulted to an unlimited sizei.e. when I want to add new operators and the screen is already full, I either need to shrink the zoom factor down to the point where I cant read the icons anymore, or I have to pick up the new operator and drag at the bottom / right side until it scrolls enough to drop it where I want. If the canvas had an unlimited size, I could simple use the scroll bars to scroll to where I want.
19.     Scroll wheel functionality to zoom in / out (and have it zoom in around the currently selected object)
20.     Option to have operations automatically resize when longer attribute names are added to them (can workaround by mimimizing and then maximizing, but would be better to have the operators automatically adjust size to fit)
21.     Ability to sort output attributes in an operator (just like table column sort functionality). If I create a new operator, it would be nice to group it with related operators instead of always having it at bottom.
22.     re: column sort functionality drag and drop instead of up/down arrows would be nice
23.     It would be nice if, when adding output attributes to (for example) the expression operator, it would choose the default data type the same way the dimension attribute wizard works (i.e. if it sees ind in name, automatically choose char(1)
24.     Change the zoom in button so that it zooms in on the currently selected operator i.e. if I click on a particular table and then hit the zoom in button, it should keep the table I selected centered
25.     When in a mapping, clicking on a line should highlight the attributes on both the source and target operators (to allow easier visual ID of where data is coming from / going to)
26.     AWM has a feature where table attributes can be sorted so that mappings between them don't have "crossed" lines. Similar functionality in OWB would be very appreciated. It may be harder to automate this, but even manual drag and drop of attributes within an operator would be nice.

if you could copy/paste this whole thing into the submission form (see feedback option under the numbers menu). The more people that request these features, the more likely we are to see them in an update or the next version.
Jason

Storing degenerate dimension info in an OWB 10gR2 cube

Hi,
I'm looking to store "degenerate dimension" information in an OWB 10gR2 cube. Typically in a DW, I'd just add a column to the fact table to hold the information. However, in OWB 10gR2, I don't see any way to add a column to a "cube" object - everything appears to either have to be a dimension surrogate key, or a measure.
Not sure the best way to proceed. We've coded all our ETL using the standard cube objects and letting OWB do all the surrogate key lookup, etc. automagically for us, so I'd hate to be forced to go back to manually loading the fact tables. Is there an easier way?
Thanks,
Scott

Hi Scott,
The problems I've run into with dimensions and cubes are:
Dimensions abort if there are many roles - about 4 is the limit. This is very bad for the time dimension which should have 100s of aliases. Unfortunately I can't use the Time dimension wizard because we require ISO-Weeks. My workaround is to create several Time dimensions manually and assign 4 roles to each.
Naming - when I create several time dimensions with 4 roles each (see above), in the BI layer, the foreign keys in cubes are always named something like time_dim, time_dim_1, etc rather than named for role names. I have to manually rename them in tables and then when I deploy BI objects, manually rename them again in Discoverer Administrator.
Redeploying just about anything is broken. Table upgrade plans are invalid (even after running the grant scripts that support recommended), Redeploying BI objects runs successfully but if you read the job output, it does nothing since the objects already exist - again, the upgrade option does nothing.
Column ordering is random when deploying objects so I've taken to sorting alphabetically for our BI objects. Not ideal by any means.
As for the time savings with lookups, I don't save nearly that much. Due to many composite keys in our source systems, I assign surrogate keys in our staging area. These become the "business keys" in our DW layer and consequently I end up doing lookups against the staging MAP tables - really no more or less difficult than looking up dimension keys. That is more a design choice than anything and may just be my "Kimball" habits dying hard.
I'm encouraged to hear that you are having success. I chose to utilize the Dimensions and Cubes and so far am sticking with it. I'm hoping that there will be patches and/or point releases to OWB soon!

OWB 10GR2 - Information

Hello,
I just want to know if the software OWB 10gR2 is still maintained by oracle (evolutionary maintenance) or in order to fix some bugs we have to upgrade to 11g ?

I guess it depends on the nature of the bug in order for you to successfully upgrade to 11g.
Cheers
David

/* + ordered */ hint used in relational cube operator

We are using OWB 10.2.0.3 build a relational datamart. The mapping which loads the fact table has a joiner joining three tables, the result of the joiner goes into the cube operator. We realized that the code created by the cube operator contains the /* + ordered */ hint. This might make sense if you do not have any joiner outside the cube operator, but in our case the hint results into a sub-optimal explain plain.
Has anyone any idea how to get rid of this hint? What is the reason for this hint?
Thanks in advance, Maren

Hi,
if you open the configuration of the mapping, select cube operators, select your target cube, select table operators, selct your cube table -> what do you see in the property 'Loading hints'?
Regards,
Detlef

OWB 10gR2 and CWM Export

Hi,
In earlier versions of OWB it was possible to do a meta data export generating a CWM file (XML).
Is there any way to generate such a file in OWB10gR2?
Is there any other way to transfer the metadata from OWB 10gR2 to query tools like Cognos or BO?
Regards
Maurice

Hello Maurice,
I don't know about CWM export in Paris and neither about Meta-Data exchange with Cognos. However, I've some information on BO. If you use the Paris to generate MOLAP cubes in the Oracle DB, you can use the BO Universe Builder provided by BO, in order to generate a BO Universe for the cube.
So you model, deploy, and load your MOLAP data mart with Paris, use the BO universe Builder for Oracle OLAP to generate the meta-data layer for BO automatically, and can start modelling nice BO reports right away.
Regards
Holger

Standard report for leading prod. order with variance value & qty.

Hi Gurus,
I am searching for a standard Tcode to get the variance value/Qty against all the leading production orders which are settled in a particular period.
Note: If a leading production order contains more than one sub orders then the qunatity of sub orders should merge with the leaing order and all the variance from individual sub orders should also merge with the leading production order.
Please suggest the tables and fileds are used for storing the leading and non leading production order and its corresponding Planned/Actual and Variance values.
Please guide me.
Regards,
Ramakant.

Hi,
Thanks for your early response.
Actually i am unable to use tcode: KKBC_ORD rather able to use KKBC only.
Using KKBC (Target/actual/production variance) and activating collective order from setting option only able to see the details for a particular leading production order only relating toa particular period .
My requirement is to display all the production orders settled in a particular period along with (Material code /order number/ Qty/Actual value and variance value).
Note: If any leading order exists then that should be collective order i.e.(value and Qty should come against the leading order number).
If there is any standard tcode exists please let me know else provide the tables and logic to bui;d up such a report.
Regards,
Ramakanta

About Using Hints in OWB

Hi ALL
I am facing one diffculty about using hints in OWB. Through the mapping configuration, i am trying to use INDEX hint for my table. The generated code is correct, but is a very redundant code. Hint is repeating multiple times in the sub query. Below is the sample:
SELECT
/*+ INDEX ("My_Table1") INDEX ("My_Table2") */
Column1,
Column2,
Column3,
Column4,
Column5,
Column6,
Column7,
Column8
FROM
( SELECT
/*+ INDEX ("My_Table1") INDEX ("My_Table2") INDEX ("My_Table1") INDEX ("My_Table2") INDEX ("My_Table1") INDEX ("My_Table2") INDEX ("My_Table1") INDEX ("My_Table2") INDEX ("My_Table1") INDEX ("My_Table2") INDEX ("My_Table1") INDEX ("My_Table2") INDEX ("My_Table1") INDEX ("My_Table2") INDEX ("My_Table1") INDEX ("My_Table2") INDEX ("My_Table1") INDEX ("My_Table2") INDEX ("My_Table1") INDEX ("My_Table2") INDEX ("My_Table1") INDEX ("My_Table2")
Column9,
Column10,
Column11,
Column12
FROM
( SELECT .....
The text under the hint ("My_Table1") INDEX ("My_Table2") is repeating again and again. Any way of controlling the code being generated..?

From a performance standpoint, using hints in that manner may cause poorer performance than without. The INDEX (table_name) hint directs the optimizer to evaluate all relevant indexes. It is very easy to build an example where your column of interest is the last column in a multicolumn index and have a query perform worse than what a full table scan would produce. Oracle's recommendation is to "use hints sparingly" and that is for good reason.

Ordered hints in Oracle:

There's a fairly popular Ordered Hint example on the web as follows:
<CODE>
select /*+ ordered use_nl(bonus) parallel(e, 4) */
e.ename,
hiredate,
b.comm
from
emp e,
bonus b
where
e.ename = b.ename
</CODE>
I would like to know what the "parallel(e, 4)" clause does. Where does the "4" come from? What is "parallel" here?
I also have another question: If I have 5 tables--T_OREGON, T_UTAH, T_VIRGINIA, T_TEXAS, and T_OKLAHOMA--lined up in a join right behind a FROM, coming in at row counts of
T_OREGON: [a lot of rows; a lot more than T_UTAH]
T_UTAH: [smaller than T_OREGON] 40550 rows
T_VIRGINIA: 14 rows
T_TEXAS: 66 rows
T_OKLAHOMA: 8 rows
like so:
...from T_OREGON or, T_UTAH ut, T_VIRGINIA va, T_TEXAS tx, T_OKLAHOMA ok...
my question is, if an Ordered hint can be used here, is it that the smallest table (in this case T_OKLAHOMA) gets placed first in the join and the rest of the table ascension in the join doesn't matter? Or is it that T_OKLAHOMA gets placed first in the join, followed by T_VIRGINIA [at 14 rows], followed by T_TEXAS, T_UTAH, and finally T_OREGON?
Thank you for any help that you can provide.

It is why my other question involved the table order
behind a FROM keyword, (when employing an Ordered
hint).
Can anybody help with that question? Thanks again.With rule based optimizer order was important.
With cost based optimizer (optimizer mode one of the all_rows, first_rows, forst_rows_n) order is not important Oracle itself chooses the order how to join tables.
If you have gathered statistics leave CBO do his work. See more about optimizer here http://download.oracle.com/docs/cd/B19306_01/server.102/b14211/optimops.htm
And if you have taken this citation out of http://articles.techrepublic.com.com/5100-22_11-1050199.html then there are at least following problems:
1) article is 4 years old
2) article doesn't say anything about downsides using this hint e.g. what if your data changes and order is not relevant anymore, quite high possibility that query is changed but hint forgotten to change, maintaining additional level of complexity where Oracle has invested many many resources to assist users enhancing CBO
3) and there is at least 1 fallacy in this article these sentences are wrong:
"When Oracle evaluates table join orders, it must consider every possible combination of tables. For example, a six-way table join has 720 (permutations of 6, or 6 * 5 * 4 * 3 * 2 * 1 = 720) possible ways that the tables can be joined together. This permutation issue becomes even more pronounced when you have more than 10 joined tables in a query: For a 15-way table join, there are over one trillion (1,307,674,368,000 to be exact) possible query permutations that must be evaluated."
(emphasis is mine - Gints)
CBO abandons each permutation as soon as the cost is higher than the smallest from previous steps. So in reality CBO is much smarter than article tries to explain.
If you really want to know how CBO works then I can only suggest to get book Cost based oracle fundamentals by Jonathan Lewis.
Gints Plivna
http://www.gplivna.eu

Does the Order of the tablesjoins in Where clause matter in ORDERED hint?

The ORDERED hint requests that the tables listed in the FROM clause of a SQL statement be joined in the order specified. But does the order of the join in "where" clause matters?
Will be there any performance difference between below query? The order of the table in the FROM clause remains the same, but there is a difference in the order of conditions in the WHERE clause
SELECT /*+ ORDERED */ a.fp_i f
FROn iw_owner.revenue_fpct b,
dw_owner.fp_dinn c,
dw_owner.nanaged_at_dinn a,
dw_owner.at_dinn_curr d,
iw_owner.na_progran e,
dw_owner.fp_dinn_curr f,
dw_owner.fpn_of_at_dinn g
WHERE a.nacc_dinn_i = b.nacc_dinn_i
AND b.fp_dinn_i = c.fp_dinn_i
AND a.acc_i = d.acc_i
AND a.acc_nacc_c = e.npn_pro_c
AND a.fp_i = f.fp_i
AND b.org_fpn_dinn_i = g.org_fpn_dinn_i(+)
AND c.fp_i IN ('JG 04')
AND b.dte_dinn_i BETWEEN '1-Apr-2011' and '30-Apr-2011'
SELECT /*+ ORDERED */ a.fp_i f
FROn iw_owner.revenue_fpct b,
dw_owner.fp_dinn c,
dw_owner.nanaged_at_dinn a,
dw_owner.at_dinn_curr d,
iw_owner.na_progran e,
dw_owner.fp_dinn_curr f,
dw_owner.fpn_of_at_dinn g
WHERE c.fp_i IN ('JG 04')
AND b.dte_dinn_i BETWEEN '1-Apr-2011' and '30-Apr-2011'
AND b.fp_dinn_i = c.fp_dinn_i
AND a.nacc_dinn_i = b.nacc_dinn_i
AND a.acc_i = d.acc_i
AND a.acc_nacc_c = e.npn_pro_c
AND a.fp_i = f.fp_i
AND b.org_fpn_dinn_i = g.org_fpn_dinn_i(+)
Thanks

Gangadhar Reddy wrote:
Does it matter the order of first 2 tables?This question makes me really wonder. It's fairly easy to see how explain changes when using hints.
Your question indicates that you do not look at explain plans?
Then why are you asking about hints in the first place?
Using hints is something you do when you want to force a specific execution plan as part of some investigation, because you know something that the optimizer does no, or you have found a bug in the optimizert.
In all cases you know exactly what you are doing.
The /*+ ORDERED */ is all about you telling the optimizer to join the tables in the order you specified them in the from clause.
As to your original question, there exists a similar ORDERED_PREDICATES hint.
Regards
Peter

Unable to use XML functions in OWB 10gR2.

Hi,
I am trying to use XMLELEMENT ,XMLATTRIBUTES and XMLFOREST function for loading into a XML file using OWB 10Gr2.
During deployment it is erroring out with an error 'identifier 'XMLELEMENT' must be declared'.
Please suggest some workaround or possible reason for this problem.
Reagrds,
PHD
Edited by: user1662077 on May 21, 2009 5:17 AM

Hi Navneet,
I have followed these steps,But beyond that there is some thing which we want to achive.
We want the xml file to have a proper format like
"<?xml version="1.0" encoding="UTF-8"?>
<Header recTyp ='01' RTA='IN200200' ISIN='IN2000020220' recDate ='20090202' stmtDate='20090202' stmtTime='232020'
totNSDLPos='100' totTrnstPos='0' TotUnlckPos='10' TotLckPos='100' totPldgPos='0' totFrzPos='9' totIdtPos='9'
NSDLTotFrzPos='1' totDtlRec='2'/>
<Details recTyp='02' dpId='IN300118'>
<BenDtl benAcctNum='12345678' type='01' subtype='02' status='01'>
<FirstHldr name='op' fatherName='' PAN='HUFSOSFNKS'>
<Address type='1' pinCode='400097' phoneNum='28490520' faxNum='9999999999'>
<Part type='1'>
</Part>
<Part type='2'>
</Part>
<Part type='3'>
</Part>
<Part type='4'>
</Part>
</Address>
</FirstHldr>.........."
On the oracle blogs it was suggested that this can be achived by using XML function in the expression operator.
But while using them,OWB is not recongnising these XML function.
Regards,
PHD

OWB 10gR2 : How to configure ctl and log locations for Sql*Loader mappings?

Hi all,
I'm using OWB 10gR2 to load data in tables with Sql*Loader mappings.
In my project I have a datafile module and an Oracle module.
When creating an sql*loader mapping in the oracle module, there is two properties for this mappings that I want to modify. The first is Control File Location and the second is Log File Location. Values for those properties are equal to the data file module location. When trying to change those values I can only chose "Use module configuration location".
Somebody knows how to configure those properties with different locations as the one of the flat file module?
What I want to do is to store the data file in one directory, and control file and log file in other directories.
Thank you for your help.
Bernard

Hi,
You're right, my problem is that the dropdown only show the location associated with the flat file location even if I have other file locations created in the design repository.
The good news is that I have found the solution to solve the problem :
1) Edit the file module and in tab "Data locations", add the locations you want to use for control file and log file.
2) Open configuration window of the mapping and then the dropdown for properties Control File Location and Log File Location show new locations
I have tested my mapping after changing those properties and it's working.
Bernard

ORDERED hint in Oracle 11gR2

Do we have any difference in 11g with the way /*+ ORDERED */ hint works, compared to 10g? Will this hint cause a plan change in 11g, compared to its plan when used in 10g (assuming no other items contribute to a change of plan)?

Welcome to the forum.
Do you mean 10.1 or 10.2 and 11.1 or 11.2?
Are you aware of the existence of http://tahiti.oracle.com?
Just test it, follow these instructions:
http://oracle-randolf.blogspot.com/2009/02/basic-sql-statement-performance.html
SQL and PL/SQL FAQ

OWB 10gR2 on DB 11g and Workflow installation

Hi all,
my customer has a data warehouse running on Oracle DB 10gR2 - 2 nodes RAC - on Linux 32 bit (Red Hat enterprise 5);
ETL is developed with OWB 10gR2 and exploits Oracle Workflow (2.6.4).
They are going to upgrade the database to 11.1 (still 2 nodes RAC) on Linux 64 bit (Red Hat enterprise 5),
and they would prefer to avoid OWB upgrade.
On Oracle Support, it seems that OWB 10gR2 is certified on DB 11.1,
so our idea is to install a OWB 10gR2 repository on DB 11.1 and then export/import MDL
from current repository (OWB 10gR2 on DB 10gR2) to the new repository (OWB 10gR2 on DB 11.1).
My first question is related to Workflow:
is it correct to install Workflow on the new DB 11.1 using the software available in the home of OWB 11g software ?
is it guaranteed thet OWB 10gR2 installed on DB 11.1 will have no problems in using such installation of OWF ?
or should we install OWF 2.6.4 on DB 11.1 using the same software that was installed on db 10gR2 ?
My second question is related to *32 bit vs 64 bit platform*;
intuitively, we should install the 64 bit version of OWB 10gR2 (and OWF) on the new plaftorm hosting DB 11g:
is ita potential issue for the OWB repository migration ?
on our pc of course we will go on using OWB Client 32 bit: are there compatibility problems?
Thanks
Silvio

Hi Silvio,
is it correct to install Workflow on the new DB 11.1 using the software available in the home of OWB 11g software ?It is better to use Oracle Workflow from DB 11g.
is it guaranteed thet OWB 10gR2 installed on DB 11.1 will have no problems in using such installation of OWF ?As I know DB 11gR1 contains also 2.6.4 Workflow release, so you won't have any problem
My second question is related to 32 bit vs 64 bit platform;
intuitively, we should install the 64 bit version of OWB 10gR2 (and OWF) on the new plaftorm hosting DB 11g:
is ita potential issue for the OWB repository migration ?IMHO it is not problem
on our pc of course we will go on using OWB Client 32 bit: are there compatibility problems?There are no any problem with 32bit OWB client and OWB repository on 64bit platform
Regards,
Oleg

OWB 10gR2 - Implementation Strategy

I am trying to find what is the best strategy to install and configure a OWB environment for the following scenario on one linux server machine:
There are three Oracle databases on this machine:
database 1 = OWB repository owned by a owb_owner schema. The repository will contain all of design metadata (Source Oracle module with source tables, Target "Oracle module A" with tables, sequences, mappings and Target "Oracle module B" with tables, sequences, mappings). The ETL process is setup to transform data in two stages: 1> Source to target A 2> Target A to target B
database 2 = It will have target schema Target_A. The contents of "Oracle Module A" from the OWB repository in database 1 needs to be deployed to this "target_A" schema
database 3 = It will have target schema Target_B. The contents of "Oracle Module B" from the OWB repository in database 1 needs to be deployed to this "target_B" schema
Do I need to have OWB repository installed in database 2 and database 3 and have control center service running in each of them to facilitate execution of ETL process?
Deployment of the objects will be launched from a windows workstation via client design center / control center manager. Do I need to have control service running from client machine to facilitate deployment ? If so, would it facilitate deployment of the objects to the above two targe schemas ?
The intent is to have a process flow in OWB metadata repository of database 1 streamline the complete ETL process tied to a location owned by OWF_MGR schema in database1.
Please advice what implementation strategy will work best for the scenario. I read strategy presented in http://download.oracle.com/docs/cd/B31080_01/doc/install.102/b28224.pdf .. BUT I need more clarity on the available options.

Hi.
you will need a unified repository in each database. The unified repository in OWB 10GR2 contains both the objects for the design repository (your mappings, tables etc.) and the control center repository (runtime and deployment information such as execution times of mappings etc.). In previous versions of OWB they were separate.
Database 1: Install the unified repository and just use it as your design repository. You will use this when you log in via Designer.
Database 2: Install unified repository together with Control Center Service and use as control center repository.
Database 3: Install unified repository together with Control Center Service and use as control center repository.
While it is possible to have the Control Center Service just run on a client machine , I ran into all sorts of problems in the past with this configuration, e.g. when you log off your client machine it shuts down the Control Center Service. From my experience this setup is not very stable.
In OWB Designer you can then set up locations and configurations to deploy your mappings to database 2 or database 3.
You will need owf_mgr on both database 2 and database 3 and you need to deploy your process flows to both databases. What you can't do is to run a process flow say from database 1 or database 2 across databases 2 and 3.
Can you explain in some more detail why you have a requirement to transform data across two target servers?
Regards
Uli

LEADING or ORDERED hint in OWB 10gR2

Similar Messages

Maybe you are looking for