Returns sample cases that are representative of the cases that are used to train the data mining model.
To use this statement, you must enable drillthrough when you create the mining model. For more information about enabling drillthrough, see CREATE MINING MODEL (DMX), SELECT INTO (DMX), and ALTER MINING STRUCTURE (DMX).
SELECT [FLATTENED] [TOP <n>] <expression list> FROM <model>.SAMPLE_CASES [WHERE <condition list>] ORDER BY <expression> [DESC|ASC]]
Optional. An integer that specifies how many rows to return.
A comma-separated list of related column identifiers.
A model identifier.
Optional. Conditions to restrict the values that are returned from the column list.
Optional. An expression that returns a scalar value.
Sample cases may be generated and may not actually exist in the training data. The returned case is representative of the specified content node.
Although the Microsoft Sequence Clustering algorithm is the only Microsoft algorithm that supports using SELECT FROM <model>.SAMPLE_CASES, third-party algorithms may also support it.
The following example returns sample cases that are used to train the Target Mail mining model. Using the IsInNode (DMX) function in the WHERE clause returns only cases that are associated with the '000000003' node. The node string can be found in the NODE_UNIQUE_NAME column of the schema rowset.
Select * from [Sequence Clustering].SAMPLE_Cases WHERE IsInNode('000000003')