Creating a master index application requires in-depth analyses of your business requirements, legacy data, and data processing requirements. After the initial analysis, you can plan and design how you will create the master index application framework and how you will customize that configuration after creating the framework. After creating the master index application, you should perform an in-depth data analysis using the tools provided. This analysis will help you define matching and standardization rules and logic. The data analysis tool is provided in the service–enabled version of Sun Master Index only, so you would need to replicate your repository–based project using the service–enabled version in order to use the tool. You also need to plan and design each physical component of the Sun Master Index project and any projects referencing the master index application files..
The following topics provide information about what you need to know and do before you begin to create a master index application.
Before creating the master index application, perform a preliminary analysis against the data that will be stored in the index database to determine the structure of the records. Analyzing your data requires reviewing the message structure of each legacy system that will share data with the master index application. The master index application does not need to store every field from every system, so you can narrow the master index application object structure to just the pertinent fields from each system. However, the master index application stores the same fields for each system. A more in-depth analysis of the field values in the legacy data occurs after the initial master index application framework is created.
Before you create the Sun Master Index project, analyze the business requirements and determine which project components will help you meet those requirements. Planning the project includes defining how each external system will share information with the master index application and how the master index application will share information with those external systems. In addition, you can incorporate master index Java methods that define how the master index application processes incoming data. Master index methods can also be used to transform the data sent from external systems into a format that can be read by the master index application.
An additional consideration is whether to integrate the master index methods into a Business Process.
Before you begin developing your master index application, make sure you have obtained the following information:
The primary object to be indexed, such as a person, customer, business, and so on
Any secondary objects, such as telephone numbers and addresses
All fields to be stored in the index for both the primary and secondary objects
The name of each field as it appears on the EDM and whether the field will be a standard text field or will be populated from a menu list (if a field will be populated from a menu list, you should also define an eight-character name for the list)
The fields that are required in order to add a record or that are required for queries
The fields that will appear on reports
The fields that must be unique to an enterprise record (in other words, they uniquely identify a child object within an enterprise record)
The fields that will be used for matching
The fields that will need to be parsed or normalized prior to matching
Any special formatting requirements, such as character types, the data type, minimum and maximum values, and field size
The fields that will appear on EDM search and search results windows
The processing codes for the source systems being integrated into the index