Matching metadata

Last published : Aug 10, 2026
The add-on enriches each table's data structure with a generic data type that provides the metadata required for matching and cleansing. These data values cannot be modified by the user. Their life cycle is fully managed by the add-on.
Properties
Definition
State
Record matching state.
Cluster ID
Identifier of the cluster. Clusters '000' to '010' are reserved by the add-on
Score (%)
Match score of the record against its pivot or golden record. When the score has not been computed it is set to '-1'.
When a suspect is considered even though its score is lower than the minimum stewardship threshold then its score is set to '-1'. This situation occurs when the add-on must keep a record in an existing cluster not matter what new computations of this record's score are (retention strategy).
Score first matching (%)
Internal score of the record computed by the first algorithm.
Score second matching (%)
Internal score of the record computed by the second algorithm.
Field score (one to many)
Score for each field participating in the matching: Score (%), Score first matching (%), Score second matching (%).
Simple matching score (%)
When simple matching is used this score is against the suspicious record. This is a transient score since the record is not yet a suspect (its state does not change). The record is just a potential suspect against the suspicious record.
Target
When the state is 'merged', this value provides the link to the pivot or golden record into which this record was merged. Except for a 'merged' record that is ignored. In this case the target still remains undefined.
Merge by
Indicates how the merge has been executed either by a user or automatically by the add-on (survivorship procedure).
Timestamp
Date and time of the last modification of any kind applied to any field of the record other than the matching metadata (except state value). When the state is updated then the timestamp is updated as well.
Was golden
Indicates that this record was previously a golden record.
List of not suspect with
List of records against which this one is not a suspect record. Takes effect for all matches over time (Multi-valued complex data type (foreign key to the record, score when the not suspect was saved). The score of each record is saved to apply the 'not suspect with threshold (%)' property of the process policy.
Auto-created
Indicates if a record is auto-created by the add-on. For example, a Golden record can be auto-created when we activate the 'Automatically create new golden' option in Matching policy.
Used surrogate fields matching
Indicates that the matching score is calculated by using surrogate field.
Group at once
Grouped:
  • Indicates that the record is in a group computed by the 'Group at once' operation.
    Cluster size:
  • Number of records in the group when the 'Group at once' operation executes.
Ongoing workflow
Identifier:
  • Technical identifier of the workflow that has been launched for the record.
  • Undefined if there is no ongoing workflow.
    Name:
  • Name of the workflow corresponding to the workflow identifier.
  • Undefined if there is no ongoing workflow.
    Timestamp:
  • Date and time of the workflow creation.
  • Undefined if there is no ongoing workflow.
    User
  • User who originated the workflow.
  • Undefined if there is no ongoing workflow.
Last process policy code
Code of the last process policy executed on the record.
Last matching policy code
Code of the last matching policy executed on the record.
Last survivorship policy code
Code of the last survivorship policy that has been executed on the record
Last operation code
In certain situations, the add-on saves an information related to the last operation applied to the record. For example, in case of an automatic merge a special code is used
Date of last operation code
Gives the last date of any operations of match or merge applied to the record
Batch operation code
Use by service 'match at once' to manage life cycle of matching applied on set of records
Cleansing meta-data
Properties Definition
Cleansing state This value indicates the record's cleansing quality level. The cleansing state can be one of four values: Undefined (null), Clean, To be fixed or Fixed. When a table is enriched with Cleansing metadata and does not run any Cleansing procedures/operations, all records have an 'Undefined' state. After cleansing procedure execution a record is moved to the 'Clean' state if no defects are found. However, when a record is found to have defects it is moved to the 'To be fixed' state. After running a cleansing operation to fix the defect, the record is changed to a 'Fixed' state.
Cleansing procedure code Code of the last cleansing procedure applied to the record.
Cleansing operation Code of last cleansing operation applied to the record.
Execution date Last execution date of the cleansing procedure or operation on the record.
User User having executed the last cleaning procedure or operation.
Field The field affected by the last execution of the cleansing procedure or operation.
Quality defect During the execution of a cleansing procedure or operation if the record has a defect, then this field is set to 'Yes', otherwise it is set to 'No'.
Merged field logging Logs data each time the field is merged into the target record.
Table 54: Metadata for matching and cleansing
Related information