SAXONICA |
The xsl:attribute
element is used to add an attribute value to an
xsl:element
element or general formatting element, or to an element created using xsl:copy
.
The attribute must be output immediately after the element, with no intervening character data.
The name of the attribute is indicated by the
name
attribute and the value by the content of the xsl:attribute
element.
The attribute name is interpreted as an attribute value template, so it may contain string expressions within curly braces. The full syntax of string expressions is outlined in XPath Expression Syntax.
The attribute value may be given either by a select
attribute or by an enclosed
sequence constructor. If the select
attribute is used and the value is a sequence, then
the items in the sequence are output space-separated.
The separator
attribute is not yet available.
For example, the following code creates a <FONT> element with several attributes:
<xsl:element name="FONT">
<xsl:attribute name="SIZE">4</xsl:attribute>
<xsl:attribute name="FACE">Courier New</xsl:attribute>
Some output text
</xsl:element>
A new attribute type
is available in XSLT 2.0. This indicates the
data type of the value of the attribute. In Saxon, the value must be a built-in type defined
in XML Schema, for example xs:integer
or xs:date
. Type annotations
are only accessible if the attribute is added to a temporary tree that specifies
validation="preserve"
. The value given to the attribute must be a string
that conforms to the rules for the data type, as defined in XML Schema.
There are two main uses for the xsl:attribute
element:
It is the only way to set attributes on an element generated dynamically using xsl:element
It allows attributes of a literal result element to be calculated using xsl:value-of.
The xsl:attribute
must be output immediately after the relevant element is generated: there must
be no intervening character data (other than white space which is ignored). Saxon outputs the closing
">" of the element start tag as soon as something other than an attribute is written to the output
stream, and rejects an attempt to output an attribute if there is no currently-open start tag. Any special
characters within the attribute value will automatically be escaped (for example, "<" will be output as
"<")
If two attributes are output with the same name, the second one takes precedence.