How Import Processes Relationships

All relationships that are imported go through two passes. This is processed in the ExtractRelationships activity.

  • The First Pass: If the import is being performed using an input map, which maps records to more than one repository, all implicit relationships are first processed. During this pass, relationships of the imported data are processed according to Merge Previous Data options. When this pass is complete, all relationships for the records are final except for the relationships specified in the CONTAINS attribute.

    When import is performed using an input map, which maps data to more than one repository, the import is said to be running in a Cross-repository Relationship mode. If data is mapped to one repository, the import is said to be running in a Simple mode.

    The following table illustrates the impact of Merge Previous Data options on relationships processing:

    Note: These factors only affect relationships to child records. Relationship to parent records are always retained.
Factors that Influence Relationship Processing During Import
Merge Previous Data Null If Blank Import Mode Data in Contains Attribute Relationships Processing
N N/A cross-repository relationship Null All implicit relationships are imported. All previous relationships are dropped. After import, only relationships that are implicitly defined in the import process will be present.
Y N/A cross-repository relationship Null All previous relationships are retained. All implicit relationships are imported. After import, union of all previous relationships and newly imported implicit relationships are present.
  • The Second Pass: In the second pass, the records are processed for explicit relationships. If the records processed in the first pass have a non-null value in the CONTAINS attribute, they are processed again.
    Note: You can control how import should handle duplicate records by setting the Duplicate Row Check parameter in the Configurator. For more information on using the Configurator, refer to TIBCO MDM Installation and Configuration.

    The default is false, that is, the import process does not allow you to import duplicate records. You can change this behavior by changing the Duplicate Row Check parameter to true.

  • Floating of Record Versions: Whenever the state of a record is being changed from draft to confirmed or unconfirmed, (for example, R3) and if there is an unconfirmed version of the same record (for example, R2), the unconfirmed version (R2) is floated to a version number subsequent (R4) to the last confirmed version. This concept is known as floating of record versions. Relationship of the floated record version is the same as that of the record version whose state was changed.
    Note: Floating of records can happen in any workflow when the record state is changed from draft to confirmed or unconfirmed.