6 Data Model Management
Model refers to a data structure that consists of well-organized business data for analysis. Data Model explicitly determines the structured data which stores persistent information in a relational database and is specified in a data modeling language.
Data Model Maintenance within the Infrastructure system facilitates you to upload the warehouse data from the Operational Systems to Database Schema using JSON (ODM)/ erwin XML file or Database Catalog.
An erwin XML file is a standard tagged XML file based on the Object Property Model that can create the required data models. You can upload the XML file by hosting it on the server and customize the update process while uploading a Business Model.
An option to upload Database.XML or JSON (ODM) files instead of erwin XML for Model Upload. In addtion, you can also upload an erwin XML and convert it to JSON (ODM). A command line utility TransformErwin.sh is provided that can run on lower environment to generate Database.XML or JSON (ODM) files from the erwin XML file, thereby saving the time taken for transforming erwin.XML to Database.XML or JSON (ODM) during the model upload process. For more information, see Command Line Utility for Transforming erwin XML to Database XML or JSON(ODM).
In case of slice, you can also use a command line utility to validate only
the updated JSONs and generate the updated JSONs for the Model Upload. This reduces the
number of the files that is required for the Model Upload. You can use the
generateSliceJson.sh
utility provided that can run on lower
environment to generate JSON (ODM) file from old
Database.XML or erwin XML and new
Database.XML or erwin XML file. For
more information, see Command Line Utility for Generating Slice JSON (ODM).
The Database Catalog feature is used to generate a business model out of the database catalog information. This can be used when a database physically exists and the business model has to be reverse-generated for OFSAA metadata references. The reverse model generation feature can be extended to RDBMS based Infodoms as well. This populates the following:
- OFSAA logical model abstraction layer, that is, the JSON files for the Infodom.
- Object registration repository.
Following are the prerequisites while working with Business Model upload:
- Buffer pool has to be available to cache the table and index data.
- The page size for the Tablespace has to be created appropriately.
Following are the Model Upload modes available in the Business Model Upload window:
Table 6-1 Fields in the Business Model Upload and their Descriptions
Field | Description |
---|---|
New |
You can upload a new business model only when you are uploading a model for the first time for the selected Information Domain. This option is not available for subsequent model uploads. JSON / erwin and DB Catalog options are available for New Model Upload. |
Incremental |
Supported incremental changes include:
The existing model details are extracted and uploaded along with the specific incremental updates. This option is available only with the subsequent model uploads and captures all the metadata pertaining to the changes in the database schema. The same can be tracked to assess the impact. The Incremental option is not supported if DB Catalog is selected for the Model Upload option. |
Rebuild |
You can re-build a model on the existing model in the database. The existing model is replaced with the current model details. This option is available with the subsequent model uploads and the current model uploaded is considered as the latest model for the selected Information Domain. Any incremental changes are considered as a ‘Rebuild’ if DB Catalog is selected as the Model Upload option. |
Sliced |
You can quickly upload the Sliced model with only the incremental changes, without merging the tables or columns of an existing model. In a Sliced Model Upload you can incrementally add new tables, add/update columns in the existing tables, and add/update primary/foreign keys in the existing model. You can also drop a column or primary/foreign key. However, dropping a table is not supported. This option is available only with the subsequent model uploads.
In sliced model upload, if the version of the Base model existing in the environment is higher than the Sliced model getting uploaded, then the columns (which are not present in the Sliced model) are not dropped. For more information, see the Model Versioning section. Sliced Model Upload compares the existing entity JSON available in the aai_dmm metadata table. Based on the checksum values:
|
Note:
To access the Import Model framework within the Infrastructure system, you (Business Analysts) must have the IBMADD (Import Business Model) Function Role mapped. To access the Data Model Upload window and add Models, you must have the DMM_ADD Function mapped to the Role and the Role (for example, DMMWRITE) must be mapped to the particular User Group. For additional information, see the My Oracle Support Document ID 2773375.1.Figure 6-1 Business Model Upload Summary window
Note:
To display the summary of the previous Model Uploads, you must have a connection pool established to access data from the database. For more information on connection pooling, see OFS AAAI Application Pack Installation & Configuration Guide available in the OHC Documentation Library.You can also search for a specific model based on the Name or Type (New / Incremental / Rebuild / Sliced) existing within the system.