You can use the PL/SQL APIs for XMLType to create XMLType tables, columns, and views; construct XMLType instances from data encoded in different character sets; and access and manipulate XMLType in various ways.

• Lazy Load of XML Data (Lazy Manifestation)
  Lazy XML loading loads rows of data only when they are requested, enhancing scalability of your applications that involve large XML documents and many concurrent users.
• XMLType Data Type Supports XML Schema
  SQL data type XMLType supports XML Schema.
• XMLType Supports Data in Different Character Sets
  You can use PL/SQL to create XMLType instances from data that is encoded in any Oracle-supported character set. To do this, you use the PL/SQL XMLType constructor or XMLType method createXML().

The PL/SQL Application Programming Interfaces (APIs) for XMLType are described.

The Document Object Model (DOM) recommended by the World Wide Web Consortium (W3C) is a universal API for accessing the structure of XML documents.

The DOM is a universal API for accessing the structure of XML documents. It was originally developed to formalize Dynamic HTML, which is used for animation, interaction, and dynamic updating of Web pages. DOM provides a language-neutral and platform-neutral object model for Web pages and XML documents. DOM describes language-independent and platform-independent interfaces to access and operate on XML components and elements. It expresses the structure of an XML document in a universal, content-neutral way. Applications can be written to dynamically delete, add, and edit the content, attributes, and style of XML documents. DOM makes it possible to create applications that work properly on all browsers, servers, and platforms.

• Oracle XML Developer's Kit Extensions to the W3C DOM Standard
  Oracle XML Developer's Kit (XDK) extends the W3C DOM API. These extensions are supported by Oracle XML DB except for those relating to client-side operations that are not applicable in the database. See the Simple API for XML (SAX) interface in the Oracle XML Developer's Kit Java and C components.
• Supported W3C DOM Recommendations
  All Oracle XML DB APIs for accessing and manipulating XML data comply with standard XML processing requirements as approved by the W3C. The PL/SQL DOM supports Levels 1 and 2 of the W3C DOM specifications.
• Difference Between DOM and SAX
  DOM is the primary generic tree-based API for XML. SAX (Simple API for XML) is the primary generic event-based programming interface between an XML parser and an XML application.

Oracle XML DB extends the Oracle Database XML development platform beyond SQL support for storage and retrieval of XML data. It lets you operate on XMLType instances using DOM in PL/SQL, Java, and C.

The default action for the PL/SQL DOM API for XMLType (DBMS_XMLDOM) does the following:

• Produce a parse tree that can be accessed by DOM APIs.

• Validate, if a DTD is found. Otherwise, do not validate.

• Raise an application error if parsing fails.

DTD validation occurs when the object document is manifested. If lazy manifestation is employed, then the document is validated when it is used.

The PL/SQL DOM API exploits a C-based representation of XML in the server and operates on XML schema-based XML instances. The PL/SQL, Java, and C DOM APIs for XMLType comply with the W3C DOM Recommendations to define and implement object-relational storage of XML data in relational or object-relational columns and as dynamic memory instances of XMLType. See Preparing XML Data to Use the PL/SQL DOM API for XMLType, for a description of W3C DOM Recommendations.

• PL/SQL DOM API Support for XML Schema
  The PL/SQL DOM API for XMLType supports XML Schema. Oracle XML DB uses annotations within an XML schema as metadata to determine the structure of an XML document and the mapping of the document to a database schema.
• Enhanced DOM Performance
  Oracle XML DB uses DOM to provide a standard way to translate data between XML and multiple back-end data sources. This eliminates the need to use separate XML translation techniques for the different data sources in your environment.

When you build applications based on Oracle XML DB, you do not need the additional components provided by Oracle XML Developer's Kit (XDK). However, you can use XDK components with Oracle XML DB to deploy a full suite of XML-enabled applications that run end-to-end.

These XDK features are particularly useful for developing XML applications based on Oracle XML DB.

• Simple API for XML (SAX) interface processing. SAX is an XML standard interface provided by XML parsers and used by procedural and event-based applications.

• DOM interface processing, for structural and recursive object-based processing.

Oracle XML Developer's Kit contains the basic building blocks for creating applications that run on a client, in a browser or a plug-in. Such applications typically read, manipulate, transform and view XML documents. To provide a broad variety of deployment options, Oracle XML Developer's Kit is available for Java, C, and C++. Oracle XML Developer's Kit is fully supported and comes with a commercial redistribution license.

Oracle XML Developer's Kit for Java consists of these components:

• XML Parsers – Creates and parses XML using industry standard DOM and SAX interfaces. Supports Java, C, C++, and the Java API for XML Processing (JAXP).

• XSL Processor – Transforms or renders XML into other text-based formats such as HTML. Supports Java, C, and C++.

• XML Schema Processor – Uses XML simple and complex data types. Supports Java, C, and C++.

• XML Class Generator, Oracle JAXB Class Generator – Automatically generate C++ and Java classes, respectively, from DTDs and XML schemas, to send XML data from Web forms or applications. Class generators accept an input file and create a set of output classes that have corresponding functionality. For the XML Class Generator, the input file is a DTD, and the output is a series of classes that can be used to create XML documents conforming with the DTD.

• XML SQL Utility – Generates XML documents, DTDs, and XML schemas from SQL queries. Supports Java.

• TransX Utility – Loads data encapsulated in XML into the database. Has additional functionality useful for installations.

• XML Pipeline Processor – Invokes Java processes through XML control files.

• XSLT VM and Compiler – Provides a high-performance C-based XSLT transformation engine that uses compiled XSL stylesheets.

• XML Java Beans – Parses, transforms, compares, retrieves, and compresses XML documents using Java components.

Create an XML schema, annotate it to map XML to SQL objects, and register the XML schema.

1. Create a standard XML schema.
2. Annotate the XML schema with definitions for the SQL objects you use.
3. Register the XML schema, to generate the necessary database mappings.

An XML schema must be registered with Oracle XML DB before it can be referenced by an XML document. When you register an XML schema, elements and attributes it declares are mapped to attributes of corresponding SQL object types within the database.

After XML schema registration, XML documents that conform to the XML schema and reference it can be managed by Oracle XML DB. Tables and columns for storing the conforming documents can be created for root elements defined by the XML schema.

An XML schema is registered by using PL/SQL package DBMS_XMLSCHEMA and by specifying the schema document and its schema-location URL. This URL is a name that uniquely identifies the registered schema within the database. It need not correspond to any real location — in particular, it need not indicate where the schema document is located.

The target namespace of the schema is another URL used in the XML schema. It specifies a namespace for the XML-schema elements and types. An XML document should specify both the namespace of the root element and the schema-location URL identifying the schema that defines this element.

When documents are inserted into Oracle XML DB using path-based protocols such as HTTP(S) and FTP, the XML schema to which the document conforms is registered implicitly, provided its name and location are specified and it has not yet been registered.

• DOM Fidelity for XML Schema Mapping
  Elements and attributes declared within an XML schema get mapped to separate attributes of the corresponding SQL object type. Other information encoded in an XML document, such as comments, processing instructions, namespace declarations and prefix definitions, and whitespace, is not represented directly.

To make existing relational and object-relational data available to your XML applications, you can create XMLType views based on it. You can then access the resulting XML data using the PL/SQL DOM API.

After you register an XML schema containing annotations that represent the mapping between XML types and SQL object types, you can create an XMLType view that conforms to the XML schema.

All DBMS_XMLDOM methods are supported by Oracle XML DB, with a few exceptions.

The DOM specifies the way elements within an XML document are used to create an object-based tree structure. It defines and exposes interfaces to manage and use the objects stored in XML documents. The DOM supports storage of documents in diverse systems.

In the DOM specification, the term "document" describes a container for many different kinds of information or data, which the DOM objectifies.

When a request such as getNodeType(myNode) is invoked, it returns myNodeType, which is the node type supported by the parent node. The following constants represent the different types that a node can adopt:

• ELEMENT_NODE

• ATTRIBUTE_NODE

• TEXT_NODE

• CDATA_SECTION_NODE

• ENTITY_REFERENCE_NODE

• ENTITY_NODE

• PROCESSING_INSTRUCTION_NODE

• COMMENT_NODE

• DOCUMENT_NODE

• DOCUMENT_TYPE_NODE

• DOCUMENT_FRAGMENT_NODE

• NOTATION_NODE

Table 11-2 shows the node types for XML and HTML, and the allowed corresponding child node types.

Document

DocumentFragment

DocumentType

EntityReference

Element

Attr

ProcessingInstruction

Comment

Text

CDATASection

Entity

Notation

Oracle XML DB DOM API for XMLType also specifies the following interfaces:

• A NodeList interface to handle ordered lists of Nodes, for example:

  • The children of a Node

  • Elements returned by method getElementsByTagName() of the element interface

• A NamedNodeMap interface to handle unordered sets of nodes, referenced by their name attribute, such as the attributes of an element.

PL/SQL function newDOMDocument constructs a DOM document handle, given an XMLType value. The resulting handle is of type DOMDocument.

A typical usage scenario for a PL/SQL application is:

1. Fetch or construct an XMLType instance

2. Construct a DOMDocument node over the XMLType instance

3. Use the DOM API to access and manipulate the XML data

   Note:

   For DOMDocument, node types represent handles to XML fragments but do not represent the data itself.

   For example, if you copy a node value, DOMDocument clones the handle to the same underlying data. Any data modified by one of the handles is visible when accessed by the other handle. The XMLType value from which the DOMDocument handle is constructed is the data, and reflects the results of all DOM operations on it.

When you change the document structure underlying a DOMDocument instance, the changes are reflected in all relevant NodeList and NamedNodeMap objects.

For example, if a DOM user gets a NodeList object containing the children of an element, and then subsequently adds more children to that element (or removes children, or modifies existing children) then those changes automatically propagate to the NodeList without additional action from the user. Likewise, changes to a node in the tree are propagated throughout all references to that node in NodeList and NamedNodeMap objects.

The interfaces: Text, Comment, and CDATASection, all inherit from the CharacterData interface.

Using PL/SQL package DBMS_XMLDOM typically involves creating DOM documents, traversing or extending the DOM tree, and creating and manipulating nodes.

Figure 11-1 Using the PL/SQL DOM API for XMLType

Description of "Figure 11-1 Using the PL/SQL DOM API for XMLType"

Examples are presented of using the PL/SQL DOM API for XMLType.

CREATE TABLE person OF XMLType;

DECLARE 
  var       XMLType; 
  doc       DBMS_XMLDOM.DOMDocument; 
  ndoc      DBMS_XMLDOM.DOMNode; 
  docelem   DBMS_XMLDOM.DOMElement;
  node      DBMS_XMLDOM.DOMNode; 
  childnode DBMS_XMLDOM.DOMNode; 
  nodelist  DBMS_XMLDOM.DOMNodelist;
  buf       VARCHAR2(2000);
BEGIN 
  var := XMLType('<PERSON><NAME>ramesh</NAME></PERSON>');

  -- Create DOMDocument handle
  doc     := DBMS_XMLDOM.newDOMDocument(var); 
  ndoc    := DBMS_XMLDOM.makeNode(doc);

  DBMS_XMLDOM.writeToBuffer(ndoc, buf);
  DBMS_OUTPUT.put_line('Before:'||buf);

  docelem := DBMS_XMLDOM.getDocumentElement(doc);

  -- Access element
  nodelist := DBMS_XMLDOM.getElementsByTagName(docelem, 'NAME');
  node := DBMS_XMLDOM.item(nodelist, 0);
  childnode := DBMS_XMLDOM.getFirstChild(node);

  -- Manipulate element
  DBMS_XMLDOM.setNodeValue(childnode, 'raj');
  DBMS_XMLDOM.writeToBuffer(ndoc, buf);
  DBMS_OUTPUT.put_line('After:'||buf);
  DBMS_XMLDOM.freeDocument(doc);
  INSERT INTO person VALUES (var);
END;
/

Before:<PERSON>
  <NAME>ramesh</NAME>
</PERSON>
 
After:<PERSON>
  <NAME>raj</NAME>
</PERSON>

SELECT * FROM person;
SYS_NC_ROWINFO$
---------------
<PERSON>
  <NAME>raj</NAME>
</PERSON>
 
1 row selected.

DECLARE
  doc   DBMS_XMLDOM.DOMDocument;
  elem  DBMS_XMLDOM.DOMElement;
  nelem DBMS_XMLDOM.DOMNode;
BEGIN
  doc := DBMS_XMLDOM.newDOMDocument;
  elem := DBMS_XMLDOM.createElement(doc, 'ELEM');
  nelem := DBMS_XMLDOM.makeNode(elem);
  DBMS_OUTPUT.put_line('Node name = ' || DBMS_XMLDOM.getNodeName(nelem));
  DBMS_OUTPUT.put_line('Node value = '|| DBMS_XMLDOM.getNodeValue(nelem));
  DBMS_OUTPUT.put_line('Node type = ' || DBMS_XMLDOM.getNodeType(nelem));
  DBMS_XMLDOM.freeDocument(doc);
END;
/

Node name = ELEM
Node value =
Node type = 1

Oracle XML DB provides abstract streams and stream-manipulation methods that you can use to handle XML nodes that are larger than 64 K bytes.

Prior to Oracle Database 11g Release 1 (11.1), each text node or attribute value processed by Oracle XML DB was limited in size to 64 K bytes. Starting with release 11.1, this restriction no longer applies.

To overcome this size limitation and allow nodes to contain graphics files, PDF files, and multibyte character encodings, the following abstract streams are available. These abstract PL/SQL streams are analogous to the corresponding Java streams. Each input stream has an associated writer, or data producer, and each output stream has an associated reader, or data consumer.

• Binary Input Stream: This provides the data consumer with read-only access to source data, as a sequential (non-array) linear space of bytes. The consumer has iterative read access to underlying source data (whatever representation) in binary format, that is, read access to source data in unconverted, "raw" format. The consumer sees a sequence of bytes as they exist in the node. There is no specification of the format or representation of the source data. In particular, there is no associated character set.

• Binary Output Stream: This provides the data producer with write-only access to target data as a sequential (non-array) linear space of bytes. The producer has iterative write access to target data in binary format, that is, write access to target data in pure binary format with no data semantics at all. The producer passes a sequence of bytes and the target data is replaced by these bytes. No data conversion occurs.

• Character Input Stream: This provides the data consumer iterative read-only access to source data as a sequential (non-array) linear space of characters, independent of the representation and format of the source data. Conversion of the source data may or may not occur.

• Character Output Stream: This provides the data producer with iterative write-only access to target data as a sequential (non-array) linear space of characters. The producer passes a sequence of characters and the target data is replaced by this sequence of characters. Conversion of the passed data may or may not occur.

Each of the input streams has the following abstract methods: open, read, and close. Each of the output streams has the following abstract methods: open, write, flush, and close. For output streams, you must close the stream before any nodes are physically written.

There are four general node-access models, for reading and writing. Each access model has both binary and character versions. Binary and character stream methods defined on data type DOMNode realize these access models.

Your application acts as the client, with the parser as its service provider. The parser mode determines whether the parser or your application drives the stream dataflow.

• For a parser in push mode, your application pushes data to the parser in an output stream, and the parser returns the result of the requested operation.

• For a parser in pull mode, your application pulls data from the parser in an input stream. Each data item in the stream is the result of a parsing event.

Each access model is described in a separate section, with an explanation of the PL/SQL functions and procedures in package DBMS_XMLDOM that operate on large nodes. The name of each subprogram reflects whether it reads (“get”) or writes (“set”) data, and whether the parser is being used in push (“push”) or pull (“pull”) mode.

For all except the get-push and set-pull access models (whether binary or character), Oracle supplies a concrete stream that you can use (implicitly). For get-push and set-pull, you must define a subtype of the abstract stream type that Oracle provides, and you must implement its access methods (open, close, and so on). For get-push and set-pull, you then instantiate your stream type and supply your stream as an argument to the access method. So, for example, you would use my_node.getNodeValueAsCharacterStream(my-stream) for get-push, but just my_node.getNodeValueAsCharacterStream() for get-pull. The latter requires no explicit stream argument, because the concrete stream supplied by Oracle is used.

In this model, you retrieve the value of a DOM node, using a parser that is in push mode. Oracle XML DB writes the node data to an output stream that the parser reads.

To read a node value in this model, your application creates a binary output stream or character output stream and passes this to Oracle XML DB. In this case, the source data is the node value. Oracle XML DB populates the output stream by adding node data to the stream. If the stream is a character output stream, then the character set, C2, is the session character set, and node data is converted, if necessary, from C1 to C2. Additionally, the data type of the node can be any that is supported by Oracle XML DB. If the node data type is not character data then the node data is first converted to character data in C2. If a binary output stream, the data type of the node must be RAW or BLOB.

The procedures of the DBMS_XMLDOM package to be used for this case are:

PROCEDURE getNodeValueAsBinaryStream (n     IN DBMS_XMLDOM.domnode, 
                                      value IN SYS.utl_BinaryOutputStream);

The application passes an implementation of SYS.utl_BinaryOutputStream into which Oracle XML DB writes the contents of the node. The data type of the node must be RAW or CLOB or else an exception is raised.

PROCEDURE getNodeValueAsCharacterStream (n     IN DBMS_XMLDOM.domnode,
                                         value IN SYS.utl_CharacterOutputStream);

The node data is converted, as necessary, to the session character set and then "pushed" into the SYS.utl_CharacterOutputStream.

The following example fragments illustrate reading the node value as binary data and driving the write methods in a user-defined subtype of SYS.utl_BinaryOutPutStream, which is called MyBinaryOutputStream:

CREATE TYPE MyBinaryOutputStream UNDER SYS.utl_BinaryOutputStream (
    CONSTRUCTOR FUNCTION MyBinaryOutputStream ()
    RETURN SELF AS RESULT,
    MEMBER FUNCTION  write (bytes IN RAW) RETURN INTEGER,
    MEMBER PROCEDURE write (bytes IN RAW, offset IN INTEGER, length IN OUT 
           INTEGER),
    MEMBER FUNCTION flush () RETURN BOOLEAN,
    MEMBER FUNCTION close () RETURN BOOLEAN);
);

-- Put code here that implements these methods
...

DECLARE
  ostream     MyBinaryOutputStream = MyBinaryOutputStream ();
  node        DBMS_XMLDOM.domnode;
  ...
BEGIN
  ...
  -- This drives the write methods in MyBinaryOutputStream,
  -- flushes the data, and closes the stream after the value has been
  -- completely written.
  DBMS_XMLDOM.getNodeValueAsBinaryStream (node, ostream);
  ...
END;

In this model, you retrieve the value of a DOM node, using a parser that is in pull mode. Oracle XML DB reads the event data from an input stream written by the parser.

To read the value of a node in this model, Oracle XML DB creates a binary input stream or character input stream and returns this to the caller. The character set, C2, of the character input stream is the current session character set. Oracle XML DB populates the input stream as the caller pulls the node data from the stream so Oracle XML DB is again the producer of the data. If the stream is a character input stream, then the node data type may be any supported by Oracle XML DB and node data, if character, is converted, if necessary, from C1 to C2. If the node data is non-character, it is converted to character in C2. If a binary input stream, the data type of the node must be RAW or BLOB.

The functions of the DBMS_XMLDOM package to be used for this case are getNodeValueAsBinaryStream and getNodeValueAsCharacterStream.

FUNCTION getNodeValueAsBinaryStream(n IN DBMS_XMLDOM.domnode) 
         RETURN SYS.utl_BinaryInputStream;

This function returns an instance of the new PL/SQL SYS.utl_BinaryInputStream that can be read using defined methods as described in the section Set-Pull Model for Large Node Handling. The node data type must be RAW or BLOB or else an exception is raised.

FUNCTION getNodeValueAsCharacterStream (n IN DBMS_XMLDOM.domnode) 
         RETURN SYS.utl_CharacterInputStream;

This function returns an instance of the new PL/SQL SYS.utl_CharacterInputStream that can be read using defined methods. If the node data is character it is converted to the current session character set. If the node data is not character data, it is first converted to character data.

Example 11-5 illustrates reading a node value as binary data in 50-byte increments:

Example 11-6 illustrates reading a node value as character data in 50-character increments:

DECLARE
  istream      SYS.utl_BinaryInputStream;
  node         DBMS_XMLDOM.domnode;
  buffer       raw(50);
  numBytes     pls_integer;
  ...
BEGIN
  ...
  istream := DBMS_XMLDOM.getNodeValueAsBinaryStream (node);
  -- Read stream in 50-byte chunks
  LOOP
    numBytes := 50;
    istream.read ( buffer, numBytes);
    if numBytes <= 0 then
       exit;
    end if;
-- Process next 50 bytes of node value in buffer
END LOOP
...
END;

DECLARE
  istream      SYS.utl_CharacterInputStream;
  node         DBMS_XMLDOM.domnode;
  buffer       varchar2(50);
  numChars     pls_integer;
  ...
BEGIN
  ...
  istream := DBMS_XMLDOM.getNodeValueAsCharacterStream (node);
-- Read stream in 50-character chunks
LOOP
   numChars := 50;
   istream.read ( buffer, numChars);
   IF numChars <= 0 then
      exit;
   END IF;
-- Process next 50 characters of node value in buffer
END LOOP 
...
END;

In this model, you set the value of a DOM node, using a parser that is in pull mode. Oracle XML DB reads the event data from an input stream written by the parser.

To write a node value in this mode, the application creates a binary input stream or character input stream and passes this to Oracle XML DB.

The character set of the character input stream, C2, is the session character set. Oracle XML DB pulls the data from the input stream and populates the node. If the stream is a character input stream, then the data type of the node may be any supported by Oracle XML DB. If the data type of the node is not character, the stream data is first converted to the node data type. If the node data type is character, then no conversion occurs, so the node data remains in character set C2. If the stream is a binary input stream, then the data type of the node must be RAW or BLOB and no conversion occurs.

The procedures of the DBMS_XMLDOM package to be used for this case are setNodeValueAsBinaryStream and setNodeValueAsCharacterStream.

PROCEDURE setNodeValueAsBinaryStream(n IN DBMS_XMLDOM.domnode,
          value IN SYS.utl_BinaryInputStream);

The application passes in an implementation of SYS.utl_BinaryInputStream from which Oracle XML DB reads data to populate the node. The data type of the node must be RAW or BLOB or else an exception is raised.

PROCEDURE setNodeValueAsCharacterStream (n IN DBMS_XMLDOM.domnode,
          value IN SYS.utl_CharacterInputStream);

The application passes in an implementation of SYS.utl_CharacterInputStream from which Oracle XML DB reads to populate the node. The data type of the node may be any valid type supported by Oracle XML DB. If it is a non-character data type, the character data read from the stream is converted to the data type of the node. If the data type of the node is either character or CLOB, then no conversion occurs and the character set of the node becomes the character set of the PL/SQL session.

Example 11-7 illustrates setting the node value to binary data produced by the read methods defined in a user-defined subtype of SYS.utl_BinaryInputStream, which is called MyBinaryInputStream:

You can use an object of type MyBinaryInputStream to set the value of a node as follows:

DECLARE
  istream     MyBinaryInputStream = MyBinaryInputStream ();
  node        DBMS_XMLDOM.domnode;
  ...
BEGIN
  ...
  -- This drives the read methods in MyBinaryInputStream
  DBMS_XMLDOM.setNodeValueAsBinaryStream (node, istream);
  ...
END;

CREATE TYPE MyBinaryInputStream UNDER SYS.utl_BinaryInputStream (
    CONSTRUCTOR FUNCTION MyBinaryInputStream ()
    RETURN SELF AS RESULT,
  MEMBER FUNCTION read () RETURN RAW,
  MEMBER PROCEDURE read (bytes IN OUT RAW, numbytes IN OUT INTEGER),
  MEMBER PROCEDURE read (bytes IN OUT RAW,
                         offset IN INTEGER,
                         length IN OUT INTEGER),
  MEMBER FUNCTION close () RETURN BOOLEAN);

In this model, you set the value of a DOM node, using a parser that is in push mode. Oracle XML DB writes the node data to an output stream that the parser reads.

To write a new node value in this mode, Oracle XML DB creates a binary output stream or character output stream and returns this to the caller. The character set of the character output stream, C2, is the current session character set. The caller pushes data into the output stream and Oracle XML DB then writes this to the Oracle XML DB Node. If the stream is a character output stream, then the data type of the node may be any type supported by Oracle XML DB. In this case, the character data is converted to the node data type. If the node data type is character, then the character set, C1, is changed to C2. No data conversion occurs. If the stream is a binary input stream, and the data type of the node must be RAW or BLOB. In this case, the stream is read without data conversion.

The procedures of the DBMS_XMLDOM package to be used for this case are setNodeValueAsBinaryStream and setNodeValueAsCharacterStream.

FUNCTION setNodeValueAsBinaryStream(n IN DBMS_XMLDOM.domnode) 
         RETURN SYS.utl_BinaryOutputStream;

This function returns an instance of SYS.utl_BinaryOutputStream into which the caller can write the node value. The data type of the node must be RAW or BLOB or else an exception is raised.

FUNCTION setNodeValueAsCharacterStream (n IN DBMS_XMLDOM.domnode) 
         RETURN SYS.utl_CharacterOutputStream;

This function returns an instance of the PL/SQL SYS.utl_CharacterOutputStream type into which the caller can write the node value. The data type of the node can be any valid Oracle XML DB data type. If the type is not character or CLOB, the character data written to the stream is converted to the node data type. If the data type of the node is character or CLOB, then the character data written to the stream is converted from PL/SQL session character set to the character set of the node

Example 11-8 illustrates setting the value of a node to binary data by writing 50-byte segments into the SYS.utl_BinaryOutputStream:

DECLARE
  ostream      SYS.utl_BinaryOutputStream;
  node         DBMS_XMLDOM.domnode;
  buffer       raw(500);
  segment      raw(50);
  numBytes     pls_integer;
  offset       pls_integer;
  ...
BEGIN
  ...
  ostream := DBMS_XMLDOM.setNodeValueAsBinaryStream (node);
  offset := 0;
  length := 500;
  -- Write to stream in 50-byte chunks
  LOOP
    numBytes := 50;
    -- Get next 50 bytes of buffer
    ostream.write ( segment, offset, numBytes);
    length := length - numBytes;
    IF length  <= 0 then
       exit;
    END IF;
  END LOOP
  ostream.close();
  ...
END;

You can use subprogram DBMS_XMLDOM.useBinaryStream to determine whether to use a character stream or a binary stream to access the content of a large node.

FUNCTION useBinaryStream (n IN DBMS_XMLDOM.domnode) RETURN BOOLEAN;

PL/SQL XSLT Processor for XMLType (DBMS_XSLPROCESSOR) follows the W3C XSLT final recommendation (REC-xslt-19991116). It provides a convenient and efficient way of applying a single XSL stylesheet to multiple documents.

The methods in PL/SQL package DBMS_XSLPROCESSOR use PL/SQL data types PROCESSOR and STYLESHEET, which are specific to the XSL Processor implementation. All DBMS_XSLPROCESSOR methods are supported by Oracle XML DB, with the exception of method setErrorLog().

You create an XSLT processor, build a STYLESHEET object from a DOM document, transform the document using the processor and the stylesheet, and use the DOM API for XMLType to manipulate the result of XSLT processing.

Figure 11-3 illustrates how to use the XSLT Processor for XMLType (DBMS_XSLPROCESSOR).

Figure 11-3 Using the PL/SQL XSLT Processor for XMLType

Description of "Figure 11-3 Using the PL/SQL XSLT Processor for XMLType"

These are the steps:

1. Construct an XSLT processor using newProcessor.

2. Use newStylesheet to build a STYLESHEET object from a DOM document.

3. Optionally, you can set parameters for the STYLESHEET object using setParams.

4. Use processXSL to transform a DOM document using the processor and STYLESHEET object.

5. Use the PL/SQL DOM API for XMLType to manipulate the result of XSLT processing.

Example 11-10 transforms an XML document using procedure processXSL. It uses the same parser instance to create two different DOM documents: the XML text to transform and the XSLT stylesheet. An XSL processor instance is created, which applies the stylesheet to the source XML to produce a new DOM fragment. A DOM node (outnode) is created from this fragment, and the node content is printed. The output DOM fragment, parser, and XSLT processor instances are freed using procedures freeDocFrag, freeParser, and freeProcessor, respectively.

DECLARE
  indoc      VARCHAR2(2000);
  xsldoc     VARCHAR2(2000);
  myParser   DBMS_XMLPARSER.parser;
  indomdoc   DBMS_XMLDOM.DOMDocument;
  xsltdomdoc DBMS_XMLDOM.DOMDocument;
  xsl        DBMS_XSLPROCESSOR.stylesheet;
  outdomdocf DBMS_XMLDOM.DOMDocumentFragment;
  outnode    DBMS_XMLDOM.DOMNode;
  proc       DBMS_XSLPROCESSOR.processor;
  buf        VARCHAR2(2000);
BEGIN
  indoc := '<emp><empno>1</empno> 
              <fname>robert</fname> 
              <lname>smith</lname>
              <sal>1000</sal>
              <job>engineer</job>
            </emp>';
  xsldoc := '<?xml version="1.0"?> 
             <xsl:stylesheet version="1.0"     
                             xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
               <xsl:output encoding="utf-8"/>
               <!-- alphabetizes an xml tree -->
               <xsl:template match="*">
                 <xsl:copy>
                   <xsl:apply-templates select="*|text()">
                     <xsl:sort select="name(.)" data-type="text" 
                               order="ascending"/>
                   </xsl:apply-templates>
                 </xsl:copy>
               </xsl:template>
               <xsl:template match="text()">
                 <xsl:value-of select="normalize-space(.)"/>
               </xsl:template>
             </xsl:stylesheet>';
  myParser := DBMS_XMLPARSER.newParser;
  DBMS_XMLPARSER.parseBuffer(myParser, indoc);
  indomdoc   := DBMS_XMLPARSER.getDocument(myParser);
  DBMS_XMLPARSER.parseBuffer(myParser, xsldoc);
  xsltdomdoc := DBMS_XMLPARSER.getDocument(myParser);
  xsl        := DBMS_XSLPROCESSOR.newStyleSheet(xsltdomdoc, '');
  proc       := DBMS_XSLPROCESSOR.newProcessor;
  --apply stylesheet to DOM document   
  outdomdocf := DBMS_XSLPROCESSOR.processXSL(proc, xsl, indomdoc);
  outnode    := DBMS_XMLDOM.makeNode(outdomdocf); 
  -- PL/SQL DOM API for XMLType can be used here
  DBMS_XMLDOM.writeToBuffer(outnode, buf);
  DBMS_OUTPUT.put_line(buf);
  DBMS_XMLDOM.freeDocument(indomdoc);
  DBMS_XMLDOM.freeDocument(xsltdomdoc);
  DBMS_XMLDOM.freeDocFrag(outdomdocf);
  DBMS_XMLPARSER.freeParser(myParser);
  DBMS_XSLPROCESSOR.freeProcessor(proc);
END;
/

<emp>
<empno>1</empno>
<fname>robert</fname>
<job>engineer</job>
<lname>smith</lname>
<sal>1000</sal>
</emp>