Friday, 12 November 2010


OBIEE 10g Metadata Dictionary

Shahed Munir

You can report from OBIEE and generate a dictionary say monthly to folder on the server and present the dictionary through OBIEE. This will help users identify fields and what subject areas they are under aswell as give you an overview of Groups and Databases that exist within OBIEE.

Generating a Metadata Dictionary

You can generate a metadata dictionary to help Oracle Business Intelligence users obtain more information about metrics or attributes for repository objects. For example, users might need to resolve issues caused by confusing metadata object names, or to obtain more details when an attribute is derived in a complicated way.

A metadata dictionary is a static set of XML documents. Each XML document describes a metadata object, such as a column, including its properties and relationships with other metadata objects. These XML documents can be viewed within the Oracle BI Presentation Services user interface, or they can be viewed directly in a browser.

Use the Administration Tool to generate a metadata dictionary for your repository.Because the dictionary does not change dynamically as repository changes are made,you must generate the dictionary periodically to update the content.

The metadata dictionary files need to be hosted on a Web server, such as Oracle HTTP Server or Apache HTTP Server. When you generate the dictionary, you can set the output location to the final location on the Web server, or to a temporary location. If you generate the dictionary in a temporary location, you must then copy the files to the location on the Web server.

Note that some large repositories can contain tens of thousands of objects. Generating a dictionary for a large repository can take a significant period of time.

To generate a metadata dictionary:

1. In the Administration Tool, open your repository in offline mode. You cannot generate a metadata dictionary in online mode.

2. Select Tools, then select Utilities.

3. Select Generate Metadata Dictionary and click Execute.

4. In the Choose Directory dialog, click Browse to locate and select the location where you want to store the dictionary. You can select a destination for your dictionary in the following ways:
■ Select a local or network location. When the dictionary is generated, a subdirectory with the same name as the repository is created in that location.
The dictionary directories and files are created in that subdirectory. For example, if you select J:\BI_DataDictionary and your repository name is demo1.rpd, the dictionary files, including the style sheets, will be located in J:\BI_DataDictionary\demo1

■ If you want to use an IIS virtual directory, you can create or select a virtual directory in IIS before you generate the dictionary. When you generate the dictionary, choose the physical directory associated with the IIS virtual directory.
5. Click OK.
6. If you did not save the files directly to a location on a Web server, copy the files over to your Web server and ensure they are accessible. Refer to the documentation for your Web server for detailed information.

Using the Calculation Wizard
The location where the metadata dictionary files can be viewed is dependent on the host name and port number of your Web server, along with the directory location where you store the files.
7. You must edit the instanceconfig.xml configuration file to enable the metadata dictionary feature in the Oracle BI Presentation Services user interface, as well as grant the appropriate privilege to your users, groups, or application roles. See

After you generate a metadata dictionary, style sheets and index files are created for that dictionary. The related style sheets (XSL files) are created and stored in a directory named xsl within the repository directory.

A name index and tree index are created and stored in the [drive]:\[path]\[repository name] root directory. The index files are associated with each other so that you can quickly switch views.

Blog Archive

About Authors

Shahed Munir

Krishna Udathu

Shahed and Krishna are Oracle / Big Data Experts