Use Nested FOR XML Queries
The xml data type and the TYPE directive in FOR XML queries enable the XML returned by the FOR XML queries to be processed on the server as well as on the client.
Processing with xml Type Variables
You can assign the FOR XML query result to an xml type variable, or use XQuery to query the result, and assign that result to an xml type variable for more processing.
DECLARE @x xml SET @x=(SELECT ProductModelID, Name FROM Production.ProductModel WHERE ProductModelID=122 or ProductModelID=119 FOR XML RAW, TYPE) SELECT @x -- Result --<row ProductModelID="122" Name="All-Purpose Bike Stand" /> --<row ProductModelID="119" Name="Bike Wash" />
You can additionally process the XML returned in the variable,
@x, by using one of the xml data type methods. For example, you can retrieve the
ProductModelID attribute value by using the value() method.
DECLARE @i int; SET @i = (SELECT @x.value('/row/@ProductModelID', 'int')); SELECT @i;
In the following example, the
FOR XML query result is returned as an xml type, because the
TYPE directive is specified in the
FOR XML clause.
SELECT ProductModelID, Name FROM Production.ProductModel WHERE ProductModelID=119 or ProductModelID=122 FOR XML RAW, TYPE,ROOT('myRoot');
This is the result:
<myRoot> <row ProductModelID="122" Name="All-Purpose Bike Stand" /> <row ProductModelID="119" Name="Bike Wash" /> </myRoot>
Because the result is of xml type, you can specify one of the xml data type methods directly against this XML, as shown in the following query. In the query, the query() method (xml Data Type) is used to retrieve the first <
row> element child of the <
SELECT (SELECT ProductModelID, Name FROM Production.ProductModel WHERE ProductModelID=119 or ProductModelID=122 FOR XML RAW, TYPE,ROOT('myRoot')).query('/myRoot/row');
This is the result:
<row ProductModelID="122" Name="All-Purpose Bike Stand" />
Returning Inner FOR XML Query Results to Outer Queries as xml Type Instances
You can write nested
FOR XML queries where the result of the inner query is returned as an xml type to the outer query. For example:
SELECT Col1, Col2, ( SELECT Col3, Col4 FROM T2 WHERE T2.Col = T1.Col ... FOR XML AUTO, TYPE ) FROM T1 WHERE ... FOR XML AUTO, TYPE;
Note the following from the previous query:
The XML generated by the inner
FOR XMLquery is added to the XML generated by the outer
The inner query specifies the
TYPEdirective. Therefore, the XML data returned by the inner query is of xml type. If the TYPE directive is not specified, the result of the inner
FOR XMLquery is returned as nvarchar(max) and the XML data is entitized.
Controlling the Shape of Resulting XML Data
Nested FOR XML queries give you more control in defining the shape of the resulting XML data. You can use nested FOR XML queries to construct XML that is partly attribute-centric and partly element-centric.
For more information about specifying both attribute-centric and element-centric XML with nested FOR XML queries, see FOR XML Query Compared to Nested FOR XML Query and Shape XML with Nested FOR XML Queries.
You can generate XML hierarchies that include siblings by specifying nested AUTO mode FOR XML queries. For more information, see Generate Siblings with a Nested AUTO Mode Query.
Regardless of which mode you use, nested FOR XML queries provide more control in describing the shape of the resulting XML. They can be used in the place of EXPLICIT mode queries.
The following topics provide examples of nested FOR XML queries.
FOR XML Query Compared to Nested FOR XML Query
Compares a single-level FOR XML query to a nested FOR XML query. This example includes a demonstration of how to specify both attribute-centric and element-centric XML as the result of the query.
Generate Siblings with a Nested AUTO Mode Query
Shows how to generate siblings by using a nested AUTO mode query
Use Nested FOR XML Queries in ASP.NET
Demonstrates how an ASPX application can use FOR XML to return XML from SQL Server.
Shape XML with Nested FOR XML Queries
Shows how to use nested FOR XML queries to control the structure of an XML document created by SQL Server.