Clusters are stored in an SQL-accessible table named

IDT_Name_MB

, where

IDT_Name

is the fully qualified name (database table name) of the IDT being clustered. Every IDT row (except Master Records) will have a corresponding membership record for each clustering strategy that it belongs to.

The table may be queried, given appropriate privileges, to join membership records with IDT records to produce a report of the clusters. Alternatively, APIs (described in the next section) may be used for this purpose.

The Membership table contains the following columns:

RECID
is reserved for internal use.
IDS_CL_ID
contains the
ClusterId
(defined by
Persistent-ID-Prefix
).
IDS_CL_NUM
contains the
ClusterNumber
.
IDS_IDT_RECID
contains the
RecID
of the corresponding IDT row.
IDS_ATTRIB
contains flags used to control the clustering process. It is a bit-field containing the following values: 0x0001 = Voting member, 0x0100 = Undecided match.
IDS_MAX_SCORE
is the score achieved by this record when it joined the cluster. The first member always has a score of 0. Note that scores are not recalculated as additional members join the cluster and, therefore, reflect the state of the cluster at a specific point in time.
IDS_SCORE_COUNT
is reserved for future use.
IDS_SCORE_ACCEPTED
is reserved for future use.
copies of the IDT Primary Key fields.