ODI Shortcut

ODI provides the shortcut feature since the 11.1.1.6.0 release.  Shortcuts look like symbolic links in Linux file system and shortcuts in windows. You keep a reference to an object that exists in another folder.

The symbolic link feature has been used widely in prepackaged apps development to handle multiple releases of the software product internally. Typically multiple top folders are created, one for each release.

Many codes can be shared from an old release to its next release.  Instead of copying the codes, a symbolic link can be created in the folder created for the new release to point to the code in the folder created for the old release.

ODI shortcut feature was created for a similar purpose.

ODI allow you to create shortcuts for these objects under the ODI Project:

• Interface
• Package
• Procedure

Shortcuts are for the codes that the developers created and modified.  It try to maximize the code reuse.  You create a code in a folder, and reference it in another folder, after you change the code in the original folder, the reference in the other folder also get the change.  You do not need to make the same change over again for each release.

I also notice that the scenarios are not shortcutable.  The scenarios (package scenario and interface scenario) are executable objects.

This dsign means that the scenarios should be regenerated after the shortcuts are created.  It also means that the scenario should be regenerated after the code in the base folder is modified.  It is kind of strange, but it is a design to help you to maximize the code reuse.  Each interface or a package is tie to an ODI model and model could be source apps version specific.  You want your code to run against a different model even the code is a shortcut.

Folders are also not shortcutable. ODI Folders are not regular folders like the way how we think in the file system directory.  Each folder carries a property called “Release Tag”.  The release tag can help redirecting the model to be used in generating the scenarios.

 

Use SQL*Loader with Named Pipe

Typically using SQL*Loader assumes that a flat file will be used as the input. The file will need to be created and generated before the SQL*Loader can take the data from the file and load the data into Oracle. The performance can be improved and the disk space can be saved if you use named pipe with SQL*Loader.

Changed Data Capture in ODI

ODI is the tool will also be used to develop the ETL (or ELT) for Oracle BI Applications. ODI supports Changed Data Capture by its journalizing module.

My understanding of CDC flow in ODI is as follows:

Aggregation for Data Warehouse, Part 2

In the aggregation for Data Warehouse, Part 1, I mentioned that there are multiple ways to do the aggregation for data warehouse. One of the method is to use the ETL tool to maintain the aggregated tables.

Knowledge Module in Oracle Data Integrator

Knowledge Module can determine how the data will be loaded from the source to the target database. Each Knowledge Module is a text based template that is used to generate the codes. It includes the skeleton, which can be in any language or technology. For example, if I select “Oracle SQL loader” LKM, ODI will generate the loader scripts. If I choose “External Table”, it generate the Insert and select statement using the external table feature.

Knowledge Module isolates the logical mapping from the different technology, so the mapping can be used from different platforms and DBMS systems. At the design time, I can specify which knowledge I want to use by the drag and drop interface. At the run time, the tool will use the Knowledge Module I specify, and dynamically generate the codes and process the job. I feel that it is very powerful and is the key component for ODI to support the hetergenious environment.

We can add our own knowledge module for the reusable components, opportunities for partnerships.

What does Forrester say about Oracle Data Integrator (Sunopsis)?

“Sunopsis’ ability to connect a wide variety of data sources and targets will allow Oracle to address the heterogeneous data management environments its enterprise architect customers must contend with.

This acquisition could be a big win for the Siebel, PeopleSoft, and JD Edwards customers who, following the acquisitions by Oracle, find themselves smack in the middle of a perhaps unwanted Oracle-dominated ecosystem.”

Please read the entire paper here:

What does Gartner say about Oracle Data Integrator (Sunopsis)?

Sunopsis fits into Oracle’s longtime strategy of building operating-system- and hardware independent tools. Fusion Middleware customers have heterogeneous IT environments, Sunopsis have a positive influence on Fusion Middleware.

See details:

Download Oracle Data Integrator (Sunopsis) from Oracle Technology Network!

Sunopsis has now renamed as Oracle Data Integrator and the evaluation version of the software is available for download here:

Learning ETL – Architecture

The central component of any ETL tool is the repository. It typically stores the metadta for all applications, including both source and target systems. It also contains the business logic about mapping and transforming the data from one to the other.

Why is Sunopsis so different?

On October 9, 2006, Oracle announced that they acquire Sunopsis.

I read the white paper from Sunopsis. I learned that the key is in their “E-LT” concept.