Principal output

When a transformation is run from the command line, the destination of the principal output is given by the -o option, and defaults to the standard output stream (that is, the console). The output is always serialized; the format of the serialization is determined by xsl:output declarations within the stylesheet itself, or by options such as !indent=yes on the command line. Command line options take precedence.

When running from an application, the destination of the principal output is controlled using the API:

• In the JAXP interface, the Transformer.transform() method takes a Result object as a parameter. For serialized output, supply a StreamResult; for tree output, supply a SAXResult, DOMResult, or StAXResult. Raw output is not possible, since JAXP was designed for XSLT 1.0.

  Saxon also supplies more specialized implementations of the Result interface such as JDOM2Writer, XOMWriter, and DOM4JWriter for writing to third party tree models.

  More generally, Saxon's internal Receiver interface implements a general-purpose push pipeline, and Receiver extends Result so any Receiver can be used as the target of a transformation.

• In the Java s9api interface, transformation output is sent to a Destination object, which has multiple implementations.

  For serialized output, supply a Serializer, which allows you to set serialization options that override the parameters in the xsl:output declaration.

  You can capture the raw output of the transformation as an XdmValue by supplying a RawDestination.

  All other options produce tree output. These include the XdmDestination for building a tree, the SAXDestination which wraps a SAX ContentHandler, and the DOMDestination for building a DOM in memory. There is also a NullDestination which discards the output, and a TeeDestination which duplicates it.

  If you want to perform further processing on the output, you can send it to a SchemaValidator for validation, or to an XsltTransformer for further transformation.

• In the C# Saxon.Api interface, transformation output is sent to an IDestination object, which has multiple implementations.

  For serialized output, supply a Serializer, which allows you to set serialization options that override the parameters in the xsl:output declaration.

  You can capture the raw output of the transformation as an XdmValue by supplying a RawDestination.

  All other options produce tree output. These include the XdmDestination for building a tree, the XmlWriterDestination which wraps a System.Xml.XmlWriter, and the DomDestination for building a DOM in memory. There is also a NullDestination which discards the output, and a TeeDestination which duplicates it.

  If you want to perform further processing on the output, you can send it to a SchemaValidator for validation, or to an XsltTransformer for further transformation.

• In the SaxonC C interface, transformation output is returned as a serialized string (pointer to a char array) or sent to a specified file.

• In the SaxonC C++, Python and PHP interfaces, there are multiple transform methods on XsltExecutable which specify how to handle the output from the different kinds of transformation invocation. These methods have the naming convention of [prefix]ReturningValue, [prefix]ReturningString and [prefix]ReturningFile (or [prefix]_returning_value, etc. in Python); where [prefix] is applyTemplates, callTemplate, or callFunction.

  The output can be returned as a serialized string (pointer to a char array) using the [prefix]ReturningString methods, or sent to a specified file using the [prefix]ReturningFile methods.

  You can capture the raw output of the transformation as an XdmValue by calling the XsltExecutable method setResultAsRawValue() (set_result_as_raw_value() in Python), and using one of the [prefix]ReturningValue methods.

  If you want to perform further processing on the output, you can send it to a SchemaValidator for validation, or to an XsltExecutable for further transformation.