Setting up partial updates correctly involves making sure that the record adapter, record manipulator, update adapter, dimension server, and other pipeline components are configured properly.
Use the following recommendations to troubleshoot a partial update pipeline:
For all incoming records, verify that your project has an property configured as the record specifier. Only one property in the project can have the
RECORD_SPECattribute set toTRUE.For the record adapter, verify that the information provided to it is in the right format. The format should correspond to the pattern that you specified in the URL field. The record adapter works properly if all of the files with the extension
*.txt.gzare formatted in the same way. For example, a URL pattern ofincoming/updates/*.txt.gzmeans that Forge reads any file that has thetxt.gzsuffix in theincoming/updatesdirectory. Each file that matches the pattern is read in sequence.If you are supplying multiple update files, verify that the Multi file field is checked in the Record Adapter editor in Developer Studio.
For the record manipulator, verify that you are correctly using the expressions used for removing and manipulating the data.
For the update adapter, verify that the Output URL field in Developer Studio points to the directory in which completed updated records from the last partial update will be placed by Forge, for the consumption by the MDEX Engine.
For the update adapter, verify that the partial update pipeline output prefix is identical to the one used for baseline updates.
For the naming format of the update source data files, use a timestamp as the naming scheme. This ensures that Forge processes the files in the proper order of their creation.
For the EAC partial update script that runs partial updates, verify that it is configured properly. The default script is created for you when you use the deployment templates to create the project; you can modify it to suit your needs.
