Applies To: SQL Server 2016 Preview

Allows the client application to browse the content of a data mining model. Client applications can use the special tree operation restrictions described at the end of this topic to navigate the content of the mining model.

The DMSCHEMA_MINING_MODEL_CONTENT rowset contains the following columns.

Column nameType indicatorLengthDescription
MODEL_CATALOGDBTYPE_WSTRThe catalog name. Microsoft SQL Server Analysis Services populates this column with the name of the database of which the model is a member.
MODEL_SCHEMADBTYPE_WSTRThe unqualified schema name. This column is not supported by Analysis Services; it always contains VT_NULL.
MODEL_NAMEDBTYPE_WSTRThe name of the model with which the content described by this row is associated.
ATTRIBUTE_NAMEDBTYPE_WSTRThe names of the attributes that correspond to this node.
NODE_NAMEDBTYPE_WSTRThe name of the node. Currently, this column contains the same value as NODE_UNIQUE_NAME, though this may change in future releases.
NODE_UNIQUE_NAMEDBTYPE_WSTRThe unique name of the node.
NODE_TYPEDBTYPE_I4The type of the node. Will generate one of the following values (3rd party data mining algorithms can extend this list):















 DM_NODE_TYPE_NN_SUBNETWORK (17) Neural network, subnetwork

 DM_NODE_TYPE_NN_INPUT_LAYER (18) Neural network, input layer (parent of input nodes)

 DM_NODE_TYPE_NN_HIDDEN_LAYER (19) Neural network, hidden layer (parent of hidden nodes)

 DM_NODE_TYPE_NN_OUTPUT_LAYER (20) Neural network, output layer (parent of output nodes)

 DM_NODE_TYPE_NN_INPUT_NODE (21) Neural network, input node

 DM_NODE_TYPE_NN_HIDDEN_NODE (22) Neural network, hidden node

 DM_NODE_TYPE_NN_OUTPUT_NODE (23) Neural network, output node

 DM_NODE_TYPE_NN_MARGINAL_STAT_NODE (24) Neural network, marginal stat node


 DM_NODE_TYPE_NB_MARGINAL_STAT_NODE (26) Neural network, marginal stat node
NODE_GUIDDBTYPE_GUIDThe node GUID. This column is not supported by Analysis Services; it always contains NULL.
NODE_CAPTIONDBTYPE_WSTRA label or a caption associated with the node. This property is primarily for display purposes.
CHILDREN_CARDINALITYDBTYPE_UI4An estimate of the number of children that the node has.
PARENT_UNIQUE_NAMEDBTYPE_WSTRThe unique name of the node's parent. NULL is returned for any nodes at the root level.
NODE_DESCRIPTIONDBTYPE_WSTRA user-friendly description of the node.
NODE_RULEDBTYPE_WSTRAn XML description of the rule that is embedded in the node.
MARGINAL_RULEDBTYPE_WSTRAn XML description of the rule that is moving to the node from the parent node.
NODE_PROBABILITYDBTYPE_R8The probability associated with this node.
MARGINAL_PROBABILITYDBTYPE_R8The probability of reaching the node from the parent node.
NODE_DISTRIBUTIONDBTYPE_HCHAPTERA table that contains the probability histogram of the node.
NODE_SUPPORTDBTYPE_R8The number of cases that support this node.
MSOLAP_MODEL_COLUMNDBTYPE_WSTRThe name of the column from the model definition to which this node pertains.
MSOLAP_NODE_SCOREDBTYPE_R8The score that was computed for this node.
MSOLAP_NODE_SHORT_CAPTIONDBTYPE_WSTRA short caption for the node that can be used for display purposes to improve readability.

The DMSCHEMA_MINING_MODEL_CONTENT rowset can be restricted on the columns listed in the following table.

Column nameType indicatorRestriction State
TREE_OPERATIONDBTYPE_UI4Optional. See below for additional remarks.

The restriction, TREE_OPERATION, is not on any particular column of the DMSCHEMA_MINING_MODEL_CONTENT rowset; rather, it specifies a tree operator. The consumer can specify a NODE_UNIQUE_NAME restriction and the tree operator (ANCESTORS, CHILDREN, SIBLINGS, PARENT, DESCENDANTS, SELF) to obtain the requested set of members. The SELF operator includes the row for the node itself in the list of returned rows. The following table describes the constants that make up the bitmap definition for the TREE_OPERATION restriction. They can be combined using the logical OR operator.


Data Mining Schema Rowsets

Community Additions