Setting Up Your Real-Time Project - 17.1

When creating a project for use in real time, note the following:

• When used for real-time processing, Name and Address, Business Data, and Empty projects can include only one input entity in the cleansing flow. Selecting multiple entities will display a message advising you that the project will not be suitable for real-time deployment. If the extra input entity will be an input to a Reference Matcher, and you will configure the project so the Reference Matcher will be the first process in your project, you can ignore this message.
• SAP projects require specific parser input fields, which correspond to fields established in the SAP application.
• You can include any process within the project flow, but be aware that several processes are designed for use only in batch mode and, due to the transactional nature of real-time projects, cannot be deployed in a real-time environment. If your project includes any of the following processes, the process will be excluded when the project is deployed for use in real time:
  • Analysis
  • Business Rules
  • Database Write
  • Decision Point
  • Dependencies
  • File Update
  • Keys
  • Load
  • Merge/Split
  • Resolve
  • Set Selection
  • Sort
  • User-defined Process
• The default Name and Address and Business Data project templates include a Relationship Linker process. When you deploy a project for real time, the deployment process automatically prepares the project to perform either a reference match or a window match, depending on how your client calls the matcher. For example, if you call the match() method, a window match is performed. If you call the matchCandidate() method, a reference match is performed. If you add a Reference Matcher process to your project and you have not set the Include Window Matching option, only reference matching is performed.
• For optimal performance, the input schema of a process should be identical to the output schema of the previous process. If the schemas do not match, your project will run but performance will be slower. If you modify a process’ input schema, you should return to the previous process and make the same updates to its output schema.
• Ensure that your schema includes all fields that are expected by your connector project. For example, when creating an SAP project, the SAP project's transformer process expects POBOX and POBOX_TEXT fields in its input schema. If you supply an input schema without these fields and try to preview the project in postcard preview, an error message is displayed. To check for required fields, open the Customer Data Parser process, open the Options menu, and check the Parser input address line types, or examine the sample trilSAP.cfg file.
• The Matching service uses an attribute named MATCHER_USER_KEY to uniquely identify records during the match process. When you deploy a project for real time, the deployment process automatically creates this attribute in the output DDL of the Relationship Linker. The default length, 100, is generally large enough to hold a unique value for each record you will send to the matcher in one call, but you can increase the length if required. Do not delete, redefine, or use this attribute for any purpose other than its default use.
• Input and output data file names and data content are ignored in a real-time environment.
• If your project contains more than 200 steps or contains steps that require a large amount of detailed information (such as US Delivery Point Validation or Asian/Pacific postal validation), it may exceed a Windows process size limitation. Edit your Windows boot.ini file to include the /3Gb switch. Refer to the Microsoft support web site for information on modifying the boot.ini file.
• The GB and US Postal Matchers include options allowing you to obtain a list of potentially matching street names and/or street details. These options are available for real time projects only. Refer to the GB and US Postal Matcher topics for information on these options.