XSLT 2.0 and XPath 2.0 Programmer's Reference, 4th Edition (288 page)

These rules have the effect that you often don't need to know whether the numbers you are dealing with are integers, decimals, or doubles. For example, if
@width
is an attribute in a schema-less document whose value is
width=“17”
, then the value of
string(@width+1)
is
18
; you never need to know that the result of the addition was actually an
xs:double
(the rules for arithmetic involving mixed types are in Chapter 8).

If you want more control over the formatting of numeric output, XSLT has a function
format-number(),
which offers detailed control. There's nothing comparable in XPath itself, but you can get rid of surplus decimal digits by using the
round-half-to-even()
function described in Chapter 13.

Examples

Changes in XPath 2.0

XPath 1.0 supported the lexical forms now used for integer literals and decimal literals, but interpreted the values as double-precision floating point. There was no support in XPath 1.0 for scientific notation.

String Literals

A
StringLiteral
represents a constant string.

Symbol	Lexical Rules
StringLiteral	( “ ([ ∧ ”])* “ )+ | ( ' ([ ∧ '])* ' )+