Schemas to Users in Oracle Physical Model

I created a relation model and sorted out the tables among the schemas. Now I am trying to create a physical model (Oracle 11g). It works and the tables are created in but there are only two Oracle users (PUBLIC & MDSYS). All the tables in the physical model have blank user (schema) field. Does it mean that I should create the users (schemas) again and sorted out the tables in the physical model?
I cannot find an option that allows me to transform the schemas in the relation model into the users in the physical model saving the links between the users and the tables.
Does this option (way) exist?
Regards,
userIV

Thank you for the prompt answer.
Anyway it seems to me strange why schemes can not be auto created ion the physical level - I think it's quite rare situation when definition in the relational model can be different from the physical one. Fortunately it's no so many users I need in mu case.

Similar Messages

Table properties doesn't propagate from Relational Model to physical model

Hi all,
I have a relationnal model with several tables and several schema.
When i create a new table in the relational model with a specified schema, the table is created on the physicla model, but the schema is not prapagate.
Is there a way to do this automatically ?
Or do i have to specify the schema on the relationnal and on the physical.
Thank you

Hi,
schema object in relational model needs to be implemented in physical model - user is the implementation of schema in Oracle physical model. User has property "implements schema". You also can omit this redirection and can assign directly user to table but in this case you'll lose the flexibility.
Here is how DDL is impacted using different option (schema_1 and table_1 in relational model):
1) physical model is not open - no implementation details
CREATE TABLE Schema_1.TABLE_1
     Column_1 NUMBER
;2) physical model is open - Schema_1 is not implemented
CREATE TABLE TABLE_1
     Column_1 NUMBER
    ) LOGGING
;3) physical model is open - Schema_1 is implemented by User1
CREATE TABLE User1.TABLE_1
     Column_1 NUMBER
    ) LOGGING
;4) physical model is open - Schema_1 is implemented by User1; Table_1 is directly assigned to User2 in physical model
CREATE TABLE User2.TABLE_1
     Column_1 NUMBER
    ) LOGGING
;Philip

Attributes For SQL Server Physical Models Aren't Saved

In version 3.0.0.665, a physical implementation implementation of data model was persisted to SQL Server 2005. The physical implementation required the use of IDENTITIES; the equivalence to an Oracle sequence.
The functionality works fine because the DDL generated correctly and the script was persisted to the target database without incidence. Both the physical model and DMD were saved. After deployment, the client required the addition of several attributes into the model. Since the model hadn't populated yet, the desired approach involved dropping the SQL SERVER database (Oracle Schema), applying the changes to the data model, generating the DDL, and creating the SQL SERVER database.
Upon inspection of the generated DDL, the IDENTITY directives were absence. Further research discovered the attributes were absence from the property sheets attached to the various physical tables. Further testing uncovered that these attributes only persist throughout the lifetime of the data modeling session.
I have not had this experience with Oracle physical models e.g.(creating trigger, sequences, etc). Most likely this is a defect in the software, but I was curious to know, if anyone else as run across this opportunity.
Warm regards,
Greg

In version 3.0.0.665, a physical implementation implementation of data model was persisted to SQL Server 2005. The physical implementation required the use of IDENTITIES; the equivalence to an Oracle sequence.
The functionality works fine because the DDL generated correctly and the script was persisted to the target database without incidence. Both the physical model and DMD were saved. After deployment, the client required the addition of several attributes into the model. Since the model hadn't populated yet, the desired approach involved dropping the SQL SERVER database (Oracle Schema), applying the changes to the data model, generating the DDL, and creating the SQL SERVER database.
Upon inspection of the generated DDL, the IDENTITY directives were absence. Further research discovered the attributes were absence from the property sheets attached to the various physical tables. Further testing uncovered that these attributes only persist throughout the lifetime of the data modeling session.
I have not had this experience with Oracle physical models e.g.(creating trigger, sequences, etc). Most likely this is a defect in the software, but I was curious to know, if anyone else as run across this opportunity.
Warm regards,
Greg

Arc is not implemented for SQL Server 2005 physical model

I added the Arc in my relational model, which is supposed to provide that references are mutually exclusive (so only one refering column can be not null). However there is not any implementation for this in the SQL Server 2005 physical model. The SQL code generated simply ignores the Arc and allows any values for the columns covered by Arc. If we check Oracle physical model. then it generates some trigger, which controls how many columns have values assigned - this is actually an implementation of Arc. But there is nothing of this kind for SQL Server model...
By the way, implementing this Arc by trigger looks like to heavy approach. The very simple table level check constraint will do the same job. But this is a different subject though.
Edited by: Otbl on Nov 3, 2011 7:34 PM

Hi Dmitry,
it's fixed in Data Modeler 3.1 EA2.
Philip

Help obtaining physical model dim level column name please

I admit I haven't read all scripting posts (yet), nonetheless someone may be gracious enough to help me with understanding the object property structure involving physical model dimension level column names.
I imported the following from an 11g instance using a dictionary import (not a DDL file import) into a virgin 3.1.1 data modeler:
create table a_dim as select 1 key, 1 payload from dual;
create dimension a_dim
level level_1 is a_dim.key
The following Transformer Rhino Javascript succinctly exhibits my ignorance:
importPackage(Packages.javax.swing);
function test_get_dim_level_column_name(o)
JOptionPane.showMessageDialog(null, model.getDesign().getRelationalDesign().getSorageDesigns().toArray()[0].getDimensionSet().toArray()[0].getLevelSet()[0]);
JOptionPane.showMessageDialog(null, model.getDesign().getRelationalDesign().getSorageDesigns().toArray()[0].getDimensionSet().toArray()[0].getLevelSet()[0].getColumnSet());
JOptionPane.showMessageDialog(null, model.getDesign().getRelationalDesign().getSorageDesigns().toArray()[0].getDimensionSet().toArray()[0].getLevelSet()[0].getColumnSet().toArray()[0]);
The first result is LEVEL_1 as one would expect but the next two results are: B1ECDFB1-FE64-F728-8AB4-A5E7F098D02D rather than KEY.
I tried prosaic attacks such as getName() and I tried the abstruse things including determining that the class of this hex stringified result is: oracle.dbtools.crest.model.design.storage.oracle.LevelOracle$InlineList
My object inspection of this thingy left me clueless.

Re: “what are you trying to do?
- In the small: learn your object model and distill my ignorance into minimalist expositions
- In the large: The current reverse engineering from Oracle physical dimensional models to the MD model level has worked for me en masse but it seems to be unsuited for incremental development. Also subviews are not supported at the MD level. My goal is to create a transformer script to engineer selected Oracle physical model dimensions to multiple blank MD models and then manually add my subject area fact tables of interest. I expect to develop dozens, if not hundreds, of MD models.
Re: “Did you read README.rtf file in xmlmetadata\doc directory?
- No, I didn’t until this morning – it seems to be new with 3.1.1 since I don’t see it in the integrated SQL Developer Data Modeler 3.1.07 – thank you.
Re: “level.getColumns() - String is equal to level.getColumnset.toString() - it should return comma separated values (usually one) representing object ID
then when you get ID you can use model.getStorageDesign().getStorageObject( ID ) to get the object - the last is written in README.rtf file.
- Using your forum remarks and reading the README.rtf I was inspired to find my sought after “KEY” column name using:
var colid = model.getDesign().getRelationalDesign().getSorageDesigns().toArray()[0].getDimensionSet().toArray()[0].getLevelSet()[0].getColumnSet().toArray()[0];
var colname = model.getDesign().getRelationalDesign().getSorageDesigns().toArray()[0].getStorageObject(colid);
JOptionPane.showMessageDialog(null,colname);
- Thank you for helping with this indirection through the object IDs.
Some remarks en passant:
1)     Nit: getSorageDesigns should be getStorageDesigns as you likely have noted
2)     The developer workflow is clumsy – i.e. having to repeatedly open/close “Custom Libraries” slows down my iterative development. It be nice if I could leave it open
3)     In the integrated SQL Developer Data Modeler 3.1.07,: alt-T(ools) / alt-D(ata) Modeler does not allow progress into submenus
Thank you Philip.
- Kevin

Oracle Physical schema password management

Hi:
After setting up an Oracle physical schema and attaching a data store, I did a lot of work reversing tables, views etc and using them in my integrations, procedures etc. I never had to look back at the physical schema until we went into production, and the IT have informed me that there is a password ageing restriction for production systems. I made provisions for other system passwords (Planning, Essbase etc) using variables. But I couldn't get my head around to doing it at the Phsyical schema level.
Given that this is a classic problem, has anyone run into this sort of issue before, and if yes, what's the best practice to deal with it?
Any help would be greatly appreciated
Thanks

Rajesh,
I see that you are using variables in Planning, essbase. On the same lines you can uses the variable as a password in the physical schema for database servers as well.
Not that this is the best practice.
But if your IT has a password ageing restriction - how tough is it for you to change it in the Topology Manager. That is the same as changing it in a database table and refreshing a variable from it but with less overhead.
Or you can change your system to use service accounts. Usage of service accounts is the best practice for things like these. Ask them - How do they run automated jobs on Unix ? Do they change the passwords for those accounts every now and then. I think the answer that you will get is "We use service accounts for that"
Are the database passwords stored in LDAP somewhere. Perhaps you can use JNDI to authenticate your database user/password from LDAP.
In the Dataserver connection properties, you will see the "JNDI Connection" checkbox. If you select it, it will change the next tab to let you add LDAP details.
Hope that helps

Users, tablespaces, sequences & synonyms disappear from physical model

I added users, tablespaces, sequences & synonyms to my physical model, then I added tablespaces to tables & indexes, and some privileges to users. After that I saved and closed Data Modeler. When I reopened the design, all this information had disappeared. However I found the users, tablespaces, synonyms and sequences from the xml-files and for some reason Data Modeler can't show them in the physical model. Has anyone else had this kind of problem?

1. I did open the physical model and that's when I noticed that some things were missing.
2. I only have one physical model in my design.
3. After reopening the design this appeared in datamodeler.log:
2011-06-27 10:32:55,566 [main] INFO DataModelerAddin - Oracle SQL Developer Data Modeler 3.0.0.665
2011-06-27 10:39:41,407 [Thread-14] ERROR XMLTransformationManager - Unable to load object from XML: O:\MyDesigns\Design1\rel\579836B5-E123AC4CCCD2\phys\D9582E4E-2ED963CB9D32\SType\seg_0\47E390DE-0671-C4B1-8428-0F45CBEE18F8.xml
java.io.FileNotFoundException: O:\MyDesigns\Design1\rel\579836B5-E123AC4CCCD2\phys\D9582E4E-2ED963CB9D32\SType\seg_0\47E390DE-0671-C4B1-8428-0F45CBEE18F8.xml
     at java.io.FileInputStream.open(Native Method)
     at java.io.FileInputStream.<init>(FileInputStream.java:106)
     at oracle.dbtools.crest.model.metadata.XMLToObjectTransformer.transformToObject(Unknown Source)
     at oracle.dbtools.crest.model.metadata.XMLTransformationManager.transformFromXMLToObject(Unknown Source)
     at oracle.dbtools.crest.model.metadata.XMLTransformationManager.openPhysicalModel(Unknown Source)
     at oracle.dbtools.crest.swingui.browser.BrowserPanel$80.run(Unknown Source)

Physical model attribute "Partition Scheme" not saved. (SS2000)

Hello,
I believe this is a bug report.
Scenario:
I have a physical model of a SQL Server database. The model contains partitioned tables, 1 partitioning function and 1 partitioning scheme. If I open the physical model properties of an index or a table, set the "Partition Scheme" attribute, save, close and re-open the model, the "Partition Scheme" attribute is lost. If I don't close out and just re-open the properties page, the attribute is retained. Some how, the attribute is lost in the persistence layer.
If there is a reason the attribute is being reset, please let me know.
Thank for your help, and all the great products you guys make,
Kurt
DataModeler v 3.1.1

Hi Kurt,
If there is a reason the attribute is being resetUnfortunately there is a bug causing the problem. Fix will be available in next patch release.
Philip

SQL Server to Oracle Converted Model missing users

I am doing a migration from SQL Server to Oracle with about 100 users. The captured model seems to be correct and all of the users are there. But, when creating the Oracle Converted Model, not all of the users are being converted. On the first attempt, the Converted Model had only 21 users, and on the next attempt, the Converted Model only had 16 users. Does anyone know what could be causing this?

Oracle Database XE does not allow more than one instance per server.
To define a primary key correct syntax is:
Create table summary(
Field NOT NULL PRIMARY KEY,
Or:
PRIMARY KEY constraint tabla_PK (field)
To generate the value of the primary key you must create a sequence before or after creating the table:
Create Sequence summary_seq
Start with 1
Increment by 1
Nocycle;
You can use the sequence in INSERT INTO or through a trigger:
With Insert Into:
Insert into Summary
Values (summary_seq.nextval, valkor1, value2 ,...);
With a trigger:
Create or replace Trigger BI_summary
Before Insert on SUMMARY
For each row
Begin
Select summary_seq.nextval into campoPK
From dual;
End;

Oracle Data Modeler commit change seg physical table

Hello,
Sometimes when I record a data model, Oracle Data Modeler (SVN) given as outgoing a chage of segment of physical table , if you committees are to change, duplication of files in the physical model is produced. Later when you open the physical model gives errors that have duplicate files
How you can correct this error? Why the segment change occurs?
Thanks

Hello,
Versión 4.0.0.833.
No pattern is occasionally when shooting in the data model. How you can correct this error? Why the segment change occurs?
Thanks

SQL Dev Data Modeller: Auto-generate Surrogate PKs in the Physical Model ?

How can I have the logical modeller allow the user to design with logical PKs, but then have surrogate primary keys be auto-generated off sequences by the modeller when it comes to create the physical model - as in conventional application design?
Without this facility, this tool is useless, IMO.
I want:
i). sequences to become the physical PKs by default, and that what were the logical PKs in the logical model, to become a unique key in the physical model.
ii). I want this set by default when generating the physical model....
iii). ....with an option to turn this off on a entity-by-entity basis (as not all tables will necessarily require such a surrogate PK; so the logical PK may remain the physical PK).

It is common practice that physical PKs in Oracle tables are defined from sequences (surrogate PKs), and that the logical PK from the entity becomes a unique key in the table.
This may not always be the case in every application out there, and some people may disagree, but it is nonetheless a needed feature.
My new Feature Request is therefore:
I would like to see the following additions to the product.
1. In the Preferences -> Data Modeler -> Model -> Logical, a flag that by default indicates whether the designer wishes to opt to enable this feature (ie; have all logical PKs converted to unique keys, and replaced by sequence nos. in the physical model). This flags needs to be there since in real life, albeit erroneously IMO, some people will choose not to opt to use this functionality.
2. On every entity created in the model, there needs to be a flag that allows to override this default option, as not every table will require a surrogate PK to be generated. Being able to (re)set a flag located on the entity properties (perhaps under 'Engineer To'), will accomplish this.
3. When Forward Engineering to the physical model from the logical, the following should happen.
ENTITY 1 ---------->TABLE 1
---------------------> P * Surrogate PK
* Attribute 1 -----> U * Column 1
* Attribute 2 -----> U * Column 2
o Attribute 3 ----------> Column 3
Here you can see,
- Attributes 1 & 2 (the logical PK) of the entity become a unique key in the table (columns 1 & 2),
- optional Attribute 3 becomes NULLable column 3,
- and a physical surrogate PK column is added (type unbounded INTEGER, PRIMARY KEY constraint added).
From entity DEPT as: (Examples based on SCOTT schema)
DEPTNO NUMBER(2) NOT NULL <-- Logical primary key on entity
DNAME VARCHAR2(14)
LOC VARCHAR2(13)
CREATE TABLE DEPT
(PK_DEPT INTEGER, -- New column becomes surrogate physical PK, driven from sequence defined later
DEPTNO NUMBER(2) NOT NULL, -- Former logical PK becomes a UK
DNAME VARCHAR2(14),
LOC VARCHAR2(13))
ALTER TABLE DEPT
ADD CONSTRAINT PK_DEPT PRIMARY KEY (PK_DEPT) USING INDEX PCTFREE 0
ALTER TABLE DEPT
ADD CONSTRAINT UKLPK_DEPTNO UNIQUE (DEPTNO) USING INDEX PCTFREE 0 -- Former logical PK becomes a UK (constraint name reflects this)
CREATE SEQUENCE PK_DEPT_SEQ
CREATE TRIGGER PK_DEPT_SEQ_TRG
BEFORE INSERT ON DEPT
FOR EACH ROW
WHEN (new.PK_DEPT IS NULL)
BEGIN
SELECT PK_DEPT_SEQ.NEXTVAL
INTO :new.PK_DEPT
FROM DUAL;
-- Or from 11g onwards, simply,
:new.PK_DEPT := PK_DEPT_SEQ.NEXTVAL;
END;
From entity EMP as:
EMPNO NUMBER(4) NOT NULL -- Logical primary key on entity
ENAME VARCHAR2(10)
JOB VARCHAR2(9)
MGR NUMBER(4)
HIREDATE DATE
SAL NUMBER(7,2)
COMM NUMBER(7,2)
DEPTNO NUMBER(2)
CREATE TABLE EMP
(PK_EMP INTEGER, -- New column becomes surrogate physical PK, driven from sequence defined later
FK_DEPT INTEGER, -- New FK to surrogate PK in DEPT table (maybe NOT NULL depending on relationship with parent)
EMPNO NUMBER(4) NOT NULL, -- Former logical PK becomes a UK
ENAME VARCHAR2(10),
JOB VARCHAR2(9),
MGR NUMBER(4),
HIREDATE DATE,
SAL NUMBER(7,2),
COMM NUMBER(7,2),
DEPTNO NUMBER(2))
ALTER TABLE EMP
ADD CONSTRAINT PK_EMP PRIMARY KEY (PK_EMP) USING INDEX PCTFREE 0
ALTER TABLE EMP
ADD CONSTRAINT FK_DEPT FOREIGN KEY (FK_DEPT) REFERENCES DEPT (PK_DEPT)
ALTER TABLE EMP
ADD CONSTRAINT UKLPK_EMPNO UNIQUE (EMPNO) USING INDEX PCTFREE 0 -- Former logical PK becomes a UK (constraint name reflects this)
CREATE SEQUENCE PK_EMP_SEQ
CREATE TRIGGER PK_EMP_SEQ_TRG
BEFORE INSERT ON EMP
FOR EACH ROW
WHEN (new.PK_EMP IS NULL)
BEGIN
SELECT PK_EMP_SEQ.NEXTVAL
INTO :new.PK_EMP
FROM DUAL;
-- Or from 11g onwards, simply,
:new.PK_EMP := PK_EMP_SEQ.NEXTVAL;
END;
[NOTE: I use PCTFREE 0 to define the index attributes for the primary & unique keys since the assumption is that they will in general not get updated, thereby allowing for the denser packing of entries in the indexes and the (albeit minor) advantages that go with it.
This is certainly always true of a sequence-driven primary key (as it is by its very nature immutable), but if the unique key is likely to be frequently updated, then this PCTFREE option could be user-configurable on a per table basis (perhaps under Table Properties -> unique Constraints).
For non-sequence-driven primary keys, this storage option could also appear under Table Properties -> Primary Key.
I notice no storage options exist in general for objects, so you may like to consider adding this functionality overall].
Associated Issues :
- Preferences, 'Naming Standard: Templates' should be updated to allow for the unique key/constraint to be called something different, thus highlighting that it comes from the logical PK. I've used 'UKLPK' in this example.
- Mark the physical PK as being generated from a sequence; perhaps a flag under Table Properties -> Primary Key.
- When Forward Engineering, if an entity exists without a logical PK, the forward engineering process should halt with a fatal error.
!!! MODERATOR PLEASE DELETE ME !!!

How to implement more than one physical models from the same model

Hello,
We have designed a relational model for our application and now, we would like to implement it for our 4 environments: Development, Test, User-Test and Real. Each one has the same logical schema but, different physical clauses (i.e.: in development environment we will not use partitioning clauses).
Is this possible?
Thank you,
Bernat Fabregat

Hello Bernat,
yes it's possible:
1) open your design and go to Tools>"RDBMS site administration"
2) add your sites there (Current design) - Development, Test, User-Test and Real; press ok when finish
3) go to "Physical models" node in browser and use "new" from pop-up menu - your sites will appear in list
4) after you set physical objects in one model (need to save it as well)l, you can clone those objects in other models - you can clone from lower or the same database version
You can consider usage of "name substitution" functionality during DDL generation if objects should have different name (prefix) in different environment - in this there will be no need to change the name of objects in models.
Philip

Difference between OBIEE Logical Modelling and physical modelling

HI.
what is the difference between logical modeling (BMM) and physical modelling(Physical layer) in obiee admin tool
please narrate it in definition format
thank you
rakesh

Hi rakesh,
Administrator tool consists of 3 layers
Physcial layer:-It is typically first layer built in the repository.Contains objects representing the physical data sources to which the BI server submits queries.It may conatin multiple data sources.The common way is to import metadat from different databases.You can define joins after import if they are not defined.
Business model and mapping layer:-It is defined as business or logical models of data and specifies the mapping between the physical and business layers.This is where the physical schemas are simplified to form the basis for the end user.We can say data is seperated into facts and dimension tables
Presentation layer:-simplifies and organizes the business model to make it easy for users to understand and query
go through the user guides for all the definition and subject
Hope it helps you.my points?
By,
Kranthi
Edited by: Kranthi.K on Sep 1, 2009 10:22 PM

Logical and Physical Models

Hi All,
I'm new to Designer. I need to produce a logical data model in Oracle Designer 10g. I have created a few tables in the RON and have used Design Editor to add these to a Server Model. I think this is giving me the physical model rather than the logical model.
Can you please tell me if I'm going about this the wrong way ? I need the logical model to confirm with the non-technical people before I produce the physical model for the techies.
Thanks

Billy wrote:
Hi,
John - yes I have already checked all this, all looks ok. Also, I'm only using the REPOS_MANAGER account for now so permissions should not be an issue.
Mark - when you say create an entity in the RON, I can create tables, how would I create an entity in the RON ? I can go to Diagrams in the RON and see my ERD.
When I installed the repository I do remember having to go to the SYS account to grant an execute on a package to the REPOS_MANAGER user because one of the REPOS_MANAGER packages required it.
Any direction would be appreciated ?
Thanks
BillyAh, but that IS the issue. You shouldn't use the repository owner account for your day to day work. Create another Oracle user and grant it privileges through the RAU. Then connect as that user to work in Designer. Oddly enough, the repository owner doesn't have the privileges that it needs - the privileges are stored in the repository, not database privileges.

Physical Model Data Type Issue

Hi,
I have created ware house tables using E-business suite tables for general ledger model and exported into physical model. I could see all my keys/IDs which were numbers have been converted into decimal.
Ex: journal_id in my oracle table is number (ex:1234). But when I exported into OBIEE and used view data option. All the IDs(data type number) data is showing as decimal format (1234.00).
Ex: If I want to see data for check number in answers, then it is showing as 112456.00 instead of 112456.
Please anyone had this kind of issue. how to resolve?
Is there any default setting which is causing this problem.how to find?
OBIEE version : 10.1.3.2
Appreciate your help.

Hi Sandeep,
When I opened the rpd in offline mode, it is saying the repository can only be opened as Read-Only. I clicked Yes for this and entered admin user name and password. But in Physical model, when I right clicked any table, I couldnt see find any place to modify. I tried with properties also. I am not able change data type as it read-only. How to do it in offline?
FYI,
I opened rpd when OC4J was stopped. Aslo I opened when OC4J is running. But both times I am not able to change datat type.
Thanks!!

Schemas to Users in Oracle Physical Model

Similar Messages

Maybe you are looking for