This step works with groups of records usually acquired from the Unification process. Utilizing predefined rules, which compare members of the same group, the overall group quality can be evaluated. Groups are defined by the value in groupIdColumn. If this value is empty (null), the record in question defines its own single member group with a group role of 'C'.

The following group qualities are defined:

• U - single member group of A quality.
• A - best quality group (records can be automatically assigned to single entity).
• M - minor differences in group (recommended for manual control).
• C - major differences in group or empty group id (recommended for manual control and correction).

<step id="simpleGroupClassifier" className="cz.adastra.cif.tasks.identify.groupClassification.SimpleGroupClassifier">
        <properties groupIdColumn="client_id" unificationRoleColumn="uni_role" roleColumn="client_group_role">
                <columnSets>
                        <columnSet>
                                <importantColumns>rc,firt_name</importantColumns>
                                <columns>last_name</columns>
                                <fuzzyColumns>city,street</fuzzyColumns>
                                <extraConditionForARole>uni_role in {'M', 'S', 'I', 'R'}</extraConditionForARole>
                                <extraConditionForMRole>uni_role in {'M', 'S', 'I'} and score_instance &lt; 101 </extraConditionForMRole>
                        </columnSet>
                </columnSets>
        </properties>
</step>