SHAPE (DMX)
Combines queries from multiple data sources into a single hierarchical table (that is, a table with nested tables), which becomes the case table for the mining model.
The complete syntax of the SHAPE command is documented in the Microsoft Data Access Components (MDAC) Software Development Kit (SDK).
Syntax
SHAPE {<master query>}
APPEND ({ <child table query> }
RELATE <master column> TO <child column>)
AS <column table name>
[
({ <child table query> }
RELATE <master column> TO <child column>)
AS < column table name>
...
]
Arguments
master query
The query returning the parent table.child table query
The query returning the nested table.master column
The column in the parent table to identify child rows from the result of a child table query.child column
The column in the child table to identify the parent row from the result of a master query.column table name
The newly appended column name in the parent table for the nested table.
Remarks
You must order the queries by the column that relates the parent table and child table, and the order should be ascending.
Examples
You can use the following example within an INSERT INTO (DMX) statement to train a model containing a nested table. The two tables within the SHAPE statement are related through the OrderNumber column.
SHAPE {
OPENQUERY([Adventure Works DW Multidimensional 2012],'SELECT OrderNumber
FROM vAssocSeqOrders ORDER BY OrderNumber')
} APPEND (
{OPENQUERY([Adventure Works DW Multidimensional 2012],'SELECT OrderNumber, model FROM
dbo.vAssocSeqLineItems ORDER BY OrderNumber, Model')}
RELATE OrderNumber to OrderNumber)
See Also
Reference
Data Mining Extensions (DMX) Data Definition Statements