xsl:sequence element was new in XSLT 2.0; it is used to construct arbitrary sequences.
It may select any sequence of nodes and/or atomic values, and essentially adds these to the result
sequence. The input may be specified either by a
select attribute, or by the instructions
contained in the
xsl:sequence instruction, or both (the
is processed first). Nodes and atomic values are included in the result sequence directly. Unlike
xsl:copy-of, no copy is made.
The most common use is to return a result from a function. For example:<xsl:function name="f:increment" as="xs:integer"> <xsl:param name="in" as="xs:integer"/> <xsl:sequence select="$in + 1"/> </xsl:function>
There are two other interesting usage scenarios. The first is copying atomic values into a tree. For example:<e> <xsl:sequence select="1 to 5"/> <br/> <xsl:sequence select="6 to 10"/> </e>
which produces the output
<e>1 2 3 4 5<br/>6 7 8 9 10</e>.
The second, more important, is constructing a sequence-valued variable. A variable
is sequence-valued if the variable binding element (e.g.
has non-empty content, an
as attribute, and no
This produces the sequence (1, 4, 9, 16, 25) as the value of the variable.
xsl:sequence instruction may be used to produce any sequence of nodes and/or
If nodes are constructed within a sequence-valued variable, they will be parentless. For example, the following code creates a variable whose value is a sequence of three parentless attributes:<xsl:variable name="seq" as="attribute() *"> <xsl:attribute name="a">10</xsl:attribute> <xsl:attribute name="b">20</xsl:attribute> <xsl:attribute name="a">30</xsl:attribute> </xsl:variable>
It is quite legitimate to have two attributes in the sequence with the same name; there is
no conflict until an attempt is made to add them both to the same element. The attributes can
be added to an element by using
<xsl:copy-of select="$seq"/> within an
xsl:element instruction or within a literal result element. At this stage the usual
rule applies: if there are duplicate attributes, the last one wins.