Distribution (clsColumn)
The Distribution property of an object of ClassType clsColumn identifies the statistical distribution of the column's data.
Data Type
String
Access
Read/write for columns with a SubClassType of sbclsRegular whose IsKey and IsParentKey properties are False, read-only for all others.
Remarks
For columns whose IsKey or IsParentKey properties are set to True and columns with a SubClassType of sbclsNested, this property returns an empty string.
Access
Read/write
Remarks
This property specifies the column's statistical distribution. This is used to optimize performance by the mining model's mining algorithm and can be left unspecified.
This property applies only to columns of SubClassType sbclsRegular, if they are not used as key or parent key columns. (That is, it applies only to regular columns whose IsKey and IsParentKey properties are False.) Supported values for this property are listed in the MINING_SERVICES schema rowset in the SUPPORTED_DISTRIBUTION_FLAGS column.
Examples
Setting the Distribution Property
The following example creates a new column and sets its Distribution property, among others:
'------------------------------------------------------------------------
' Add a new column to the mining model called Unit Sales and relate
' this column to the Sales cube measure of the same name. Set the
' column's data type to Integer, and identify the data content in it as
' being continous and logarithmically normalized. Finally, identify this
' column as being predictable.
' Assume the existence of a DSO Cube object, dsoCb.
'------------------------------------------------------------------------
'Add another column to the model.
Set dsoColumn = dsoDmm.Columns.AddNew("Unit Sales")
'Identify this column as being based on the Unit Sales measure.
Set dsoColumn.SourceOlapObject = dsoCb.Measures("Unit Sales")
'Identify the column type.
dsoColumn.DataType = adInteger
'Identify this column's content as being continuous.
dsoColumn.ContentType = "CONTINUOUS"
'Identify the statistical distribution of this data.
dsoColumn.Distribution = "LOG_NORMAL"
'Identify the column as being predictable.
dsoColumn.IsPredictable = True