repository building

[Former Member]

Hi all,

I have been given a logical&physical data model, Now I need to build a repository.

Could any one pls direct me on how i need to start and proceed further.

Should I use the buisness content and edit the tables, create new tables,fields if necessary....... is that the normal process?

or should i ask for the scripts created in erwin, can I use that scipts to build a repository automatically? yes, How ?

regards,

Ve

[former_member182007 · Accepted solution]

Veena

This is totally depends on your business requirment as in how to create or design the repository. Here If you unarchive the standard business contents all the default tables and their respective fields with properties as well ill be coming.

These are the standard repository and you can also modify it as per your requirments as in by adding or delting fields and by modifiying the properties.

Non standard repositories need to be designed from scratch.Since there is no functionality in the Console that checks a repository for potential performance problems, a system administrator has all the freedom in the definition of a repository, which could lead later on to performance problems. Many times these problems are discovered at a very late stage

of an implementation, often when it comes to loading large bulk of data into the system or syndicating these data back to

other systems, just to name a few situations. Hence it becomes necessary to design MDM repositories ,

which optimize system performance.

A well designed repository should have these features:

1. Contains a limited number of tables and fields

2. Is easy to navigate in the MDM Data Manager

3. Performs well during search & maintenance of records

4. Shows good performance during Import load & syndication

5. Is free of workarounds

6. Has the data in the right context

Maximum records should be in main table.

Minimal records should be in lookups.

In Languages

For standard repositories, it is advisable to delete the additional not necessary languages. For non standard ones, only required languages must be activated. Languages can always be added later on as required.

Lookup Tables

Keep the Number of Records and Fields Limited.

Qualified Lookup Tables

Keep the Number of Records Limited. Choose the Right Non-Qualifiers.More non-qualifiers more the no. of records.

Index

MDM spends extra memory and processing time to maintain sort indices. Indexing can take a lot of time during repository load.

Having larger repositories with numerous sorted fields can have a dramatic impact on the system performance.

Loading a repository with a limited number of records with the update indices option also gives you a pretty good indicator about the repository design.

moreover you can also discuss with your business to make your Repository in a good shape.

Hope this info Help.

Deep

[Former Member · Accepted solution]

Jignesh/Deepak, thanks for your quick replies.

now I already have the logical and phical model with me, how can i make use of that. As you know using erwing tool you

can create scripts, which can be run to create a data base according to the physical model.(my conceptual knowledge).

can we do something like that in SAP MDM ?

or do you think that i just need to use the data model for refrence to create/delete the contents in my business content repositoy.

regards,

ve

[Former Member · Accepted solution]

>>can we do something like that in SAP MDM

No, this can not be done in MDM.

Please understand that MDM is not a database, it is a wrapper on top of a database (same way as SAP is a wrapper on top of database). So whatever you do must be done in MDM, not directly in the database.

Once you have the physical/logical data model, you need to decide how you will model MDM repository, if you can re-use standards rep. with minor modifications, then use that otherwise you need to create whole rep. from scratch, your MDM knowledge will come into play here. Logical/Physical data model won't tell you if you need to create Flat look-up or hierarchical look-up, its upon you to decide the best option.

As far as i know, there are no plug-in in Erwin for MDM (they do have plug-ins for other databases, but MDM is not a database anyways).

Thanks

Aamir

[former_member182007 · Accepted solution]

In addition to above posts. it is always up to you and your business to define a Data Model for Repository which suits your requirments.

Nothing here can b reflected from Database, as MDM is a Table Driven structure. Now to make a Shape for Model either you can use SAP ready contentents or make your own.

Hope it helps.

Deep

[Former Member · Accepted solution]

Hi Veena

This is not possible as the data types supported in SAP MDM are different then DB level data types. We have some very special tables like Taxonomy, Qualified tables and tuples supported in MDM for better data modeling.

Also the metadata of mdm will be different from that of Oracke. Reason is mdm builds the structure according to predetermined data types and not all the oracle data types are supported by mdm.

Since you already have the logical and physical data model with you the next step is to do a MDM Data modeling and prepare an MDM data model from the above two models you have. For this we need to understand more about the different MDM data types and how they can be utilized to full potential. You can also start with the Std business content repository, if the target system is SAP based, and delete/add tables/fields depending on the requirements.

hope this helps-Ravi

[Former Member · Accepted solution]

Hi Veena,

I Agree with Aamir it is not possible to create script which can be run to create a data base accoreding to the physical model

I would suggest utilize SAP standered buisness content reposity and modify the chages for e.g. create/Delete content of repository as per your buisness requirement

Thanks,

Jignesh Patel

[Former Member · Accepted solution]

All, thanks for all your responses...... still have a dbt

My data model has relationships, (like one to many, many to one), does any of this indicate which type of table(qualified, hierchy, etc) to be used. So, how to decide these tables based on the relationships.

pls clarify

thanks,

ve

[former_member206388 · Accepted solution]

Hi Veena.

There are 2 types of relationship which MDM suports.

1. Sibling - Group of main table records which are equivalent to each other (Ex:Related Items, Substiute Items )

2. Parent/Child - Group of main table records which are not equivalent in which one item is a parent and others are child records (Ex:Bill of Materials, Bundle Items, Kits etc. )

- Sibling types can be utilsed using Main table

- Parent/Child are supported with Flat, Qualifier, Hiearchy and Main table (Not Taxonomy).

Based on your business scenario, you can find which one to use.

Regards

Bala

[Former Member · Accepted solution]

Hi Veena,

Details as follows:

One to Many Relationship: in this case you can use Hierarchy Table

Many to One : in this case you can use Qualified or Tuple Table

Note : it's depend up on your business requirement how u can utilize this table

Thanks,

Jignesh Patel

[Former Member · Accepted solution]

Hi Every one,

Irrespective of materials/cust/vend/products data b/w different organisation, i guess there would some basic format of designing the data modal.

Like for materials, the following data init should be maintained in subtables and some in qualified tables.

I guess this is nothing but MDM Bussiness content.

So incase if you are designing repository for the first time it is good to use the BC and modify it(add-or delete fields) if required.

Experts please evaluate this, if i am wrong

KR

John

[Former Member · Accepted solution]

Hi All,

Thanks a lot.

your responses were of great help to me.

I wish I had more points to give to everyone who contributed.

Thanks to SDN.

Rgds,

Ve

[Former Member]

Also, If I am using businesscontent, How to decide which table will be hierchy, flat, taxonomy, qualified, tuples.

pls .... share your experience.

thanks in advance,

ve

[Former Member]

Hi Veena,

SAP provide business content and standard repository you can utilize

kindly note Following point

1. Taxonomy Table : this table generally used for classification data of SAP ECC System i.e. attributes and

attributes value which is dynamically changing

2. Flat Table : this table allowed to select lookup value in SAP its Check table which allow to select correct lookup value

3. Qualified Table : this is special table in which use to maintain more information abt value need to maintain mostly used for subtable

for e.g. in SAP Plant IM11, IM12, IM13 contain more information like storage location, MRP controller, profit Center etc for specific plant

in this value of fields information is static

4. Hierarchy : this table is for parents and child relationship for e.g. if you more then 15000 value so its difficult to search the value and all this

value are depend upon category so define this value under category so it will provide easy to search the value

5. Tuple : its a special table utilized for nested value (sub-sub table) to over come the problem of qualified table

Note : above mention point it is an example which may change depend upon your business requirement

Thanks,

Jignesh Patel