Difference between revisions of "XQuery 3.0"

From BaseX Documentation
Jump to navigation Jump to search
(5 intermediate revisions by 2 users not shown)
Line 1: Line 1:
This article is part of the [[XQuery|XQuery Portal]].
+
This article is part of the [[XQuery|XQuery Portal]]. It provides a summary of the most important features of the [http://www.w3.org/TR/xquery-30/ XQuery 3.0] Recommendation.
It summarizes the most interesting features of the
 
[http://www.w3.org/TR/xquery-30/ XQuery 3.0] Recommendations.
 
XQuery 3.0 is fully supported by BaseX.
 
  
 
=Enhanced FLWOR Expressions=
 
=Enhanced FLWOR Expressions=
  
Most clauses of FLWOR expressions can now be specified in an arbitrary order: additional {{Code|let}} and {{Code|for}} clauses can be put after a {{Code|where}} clause, and multiple {{Code|where}}, {{Code|order by}} and {{Code|group by}} statements can be used. This means that many nested loops can now be rewritten to a single FLWOR expression.
+
Most clauses of FLWOR expressions can be specified in an arbitrary order: additional {{Code|let}} and {{Code|for}} clauses can be put after a {{Code|where}} clause, and multiple {{Code|where}}, {{Code|order by}} and {{Code|group by}} statements can be used. This means that many nested loops can now be rewritten to a single FLWOR expression.
  
 
'''Example:'''  
 
'''Example:'''  
Line 50: Line 47:
 
</pre>
 
</pre>
  
In contrast to the relational GROUP BY statement, the XQuery counterpart
+
In contrast to the relational GROUP BY statement, the XQuery counterpart concatenates the values of all non-grouping variables that belong to a specific group. In the context of our example, all nodes in <code>//people/person</code> that belong to the <code>preferred</code> partition are concatenated in <code class="brush:xquery">$ppl</code> after grouping has finished. You can see this effect by changing the return statement to:
concatenates the values of all non-grouping variables that belong to a specific group.
 
In the context of our example, all nodes in <code>//people/person</code> that belong to the <code>preferred</code> partition are concatenated in <code class="brush:xquery">$ppl</code> after grouping has finished.
 
You can see this effect by changing the return statement to:
 
  
 
<pre class="brush:xquery">  
 
<pre class="brush:xquery">  
Line 102: Line 96:
 
==count==
 
==count==
  
A new {{Code|count}} clause enhances the FLWOR expression with a variable that enumerates the iterated tuples.
+
The {{Code|count}} clause enhances the FLWOR expression with a variable that enumerates the iterated tuples.
  
 
<pre class="brush:xquery">
 
<pre class="brush:xquery">
Line 176: Line 170:
 
</pre>
 
</pre>
  
A map operator is defined to be part of a path expression, which may now mix path and map operators. In contrast to the path operator, the results of the map operator will not be made duplicate-free and returned in document order.
+
In contrast to path expressions, the results of the map operator will not be made duplicate-free and returned in document order.
  
 
=Try/Catch=
 
=Try/Catch=
Line 239: Line 233:
 
=Expanded QNames=
 
=Expanded QNames=
  
A ''QName'' can now be directly prefixed with the letter "Q" and a namespace URI in the [http://www.jclark.com/xml/xmlns.htm Clark Notation].
+
A ''QName'' can be prefixed with the letter "Q" and a namespace URI in the [http://www.jclark.com/xml/xmlns.htm Clark Notation].
  
 
'''Examples:'''
 
'''Examples:'''
Line 247: Line 241:
 
=Namespace Constructors=
 
=Namespace Constructors=
  
New namespaces can now be created via so-called 'Computed Namespace Constructors'.
+
New namespaces can be created via so-called 'Computed Namespace Constructors'.
  
 
<pre class="brush:xquery">  
 
<pre class="brush:xquery">  
Line 263: Line 257:
 
=External Variables=
 
=External Variables=
  
Default values can now be attached to external variable declarations. This way, an expression can also be evaluated if its external variables have not been bound to a new value.
+
Default values can be attached to external variable declarations. This way, an expression can also be evaluated if its external variables have not been bound to a new value.
  
 
<pre class="brush:xquery">  
 
<pre class="brush:xquery">  
Line 272: Line 266:
 
=Serialization=
 
=Serialization=
  
[[Serialization|Serialization ]]parameters can now be defined in XQuery expressions. Parameters are placed in the query prolog and need to be specified as option declarations, using the <code>output</code> prefix.
+
[[Serialization|Serialization ]]parameters can be defined within XQuery expressions. Parameters are placed in the query prolog and need to be specified as option declarations, using the <code>output</code> prefix.
  
 
'''Example:'''  
 
'''Example:'''  
Line 287: Line 281:
 
=Context Item=
 
=Context Item=
  
The context item can now be specified in the prolog of an XQuery expression:
+
The context item can be specified in the prolog of an XQuery expression:
  
 
'''Example:'''  
 
'''Example:'''  

Revision as of 08:45, 30 November 2017

This article is part of the XQuery Portal. It provides a summary of the most important features of the XQuery 3.0 Recommendation.

Enhanced FLWOR Expressions

Most clauses of FLWOR expressions can be specified in an arbitrary order: additional let and for clauses can be put after a where clause, and multiple where, order by and group by statements can be used. This means that many nested loops can now be rewritten to a single FLWOR expression.

Example:

for $country in db:open('factbook')//country
where $country/@population > 100000000
for $city in $country//city[population > 1000000]
group by $name := $country/name[1]
count $id
return <country id='{ $id }' name='{ $name }'>{ $city/name }</country>

group by

FLWOR expressions have been extended to include the group by clause, which is well-established in SQL. group by can be used to apply value-based partitioning to query results:

XQuery:

 
for $ppl in doc('xmark')//people/person  
let $ic := $ppl/profile/@income
let $income :=  if($ic < 30000) then
                   "challenge" 
                else if($ic >= 30000 and $ic < 100000) then 
                   "standard" 
                else if($ic >= 100000) then 
                   "preferred" 
                else 
                   "na"  
group by $income
order by $income
return element { $income } { count($ppl) }

This query is a rewrite of Query #20 contained in the XMark Benchmark Suite to use group by. The query partitions the customers based on their income.

Result:

<challenge>4731</challenge>
<na>12677</na>
<preferred>314</preferred>
<standard>7778</standard>

In contrast to the relational GROUP BY statement, the XQuery counterpart concatenates the values of all non-grouping variables that belong to a specific group. In the context of our example, all nodes in //people/person that belong to the preferred partition are concatenated in $ppl after grouping has finished. You can see this effect by changing the return statement to:

 
...
return element { $income } { $ppl }

Result:

<challenge>
  <person id="person0">
    <name>Kasidit Treweek</name>
    …
  <person id="personX">
    …
</challenge>

Moreover, a value can be assigned to the grouping variable. This is shown in the following example:

XQuery:

let $data :=
  <xml>
    <person country='USA' name='John'/>
    <person country='USA' name='Jack'/>
    <person country='Germany' name='Johann'/>
  </xml>
for $person in $data/person
group by $country := $person/@country/string()
return element persons {
  attribute country { $country },
  $person/@name ! element name { data() }
}

Result:

<persons country="USA">
  <name>John</name>
  <name>Jack</name>
</persons>
<persons country="Germany">
  <name>Johann</name>
</persons>

count

The count clause enhances the FLWOR expression with a variable that enumerates the iterated tuples.

for $n in (1 to 10)[. mod 2 = 1]
count $c
return <number count="{ $c }" number="{ $n }"/>

allowing empty

The allowing empty provides functionality similar to outer joins in SQL:

for $n allowing empty in ()
return 'empty? ' || empty($n)

window

Window clauses provide a rich set of variable declarations to process sub-sequences of iterated tuples. An example:

for tumbling window $w in (2, 4, 6, 8, 10, 12, 14)
    start at $s when fn:true()
    only end at $e when $e - $s eq 2
return <window>{ $w }</window>

More information on window clauses, and all other enhancements, can be found in the specification.

Function Items

One of the most distinguishing features added in XQuery 3.0 are function items, also known as lambdas or lambda functions. They make it possible to abstract over functions and thus write more modular code.

Examples:

Function items can be obtained in three different ways:

  • Declaring a new inline function:
    let $f := function($x, $y) { $x + $y }
    return $f(17, 25)

    Result: 42

  • Getting the function item of an existing (built-in or user-defined) XQuery function. The arity (number of arguments) has to be specified as there can be more than one function with the same name:
    let $f := math:pow#2
    return $f(5, 2)

    Result: 25

  • Partially applying another function or function item. This is done by supplying only some of the required arguments, writing the placeholder ? in the positions of the arguments left out. The produced function item has one argument for every placeholder.
    let $f := fn:substring(?, 1, 3)
    return (
      $f('foo123'),
      $f('bar456')
    )

    Result: foo bar

Function items can also be passed as arguments to and returned as results from functions. These so-called Higher-Order Functions like fn:map and fn:fold-left are discussed in more depth on their own Wiki page.

Simple Map Operator

The simple map operator ! provides a compact notation for applying the results of a first to a second expression: the resulting items of the first expression are bound to the context item one by one, and the second expression is evaluated for each item. The map operator may be used as replacement for FLWOR expressions:

Example:

(: Simple map notation :)
(1 to 10) ! element node { . },
(: FLWOR notation :)
for $i in 1 to 10
return element node { $i }

In contrast to path expressions, the results of the map operator will not be made duplicate-free and returned in document order.

Try/Catch

The try/catch construct can be used to handle errors at runtime:

Example:

try {
  1 + '2'
} catch err:XPTY0004 {
  'Typing error: ' || $err:description
} catch * {
  'Error [' || $err:code || ']: ' || $err:description
}

Result: Typing error: '+' operator: number expected, xs:string found.

Within the scope of the catch clause, a number of variables are implicitly declared, giving information about the error that occurred:

  • $err:code error code
  • $err:description: error message
  • $err:value: value associated with the error (optional)
  • $err:module: URI of the module where the error occurred
  • $err:line-number: line number where the error occurred
  • $err:column-number: column number where the error occurred
  • $err:additional: error stack trace

Switch

The switch statement is available in many other programming languages. It chooses one of several expressions to evaluate based on its input value.

Example:

for $fruit in ("Apple", "Pear", "Peach")
return switch ($fruit)
  case "Apple" return "red"
  case "Pear"  return "green"
  case "Peach" return "pink"
  default      return "unknown"

Result: red green pink

The expression to evaluate can correspond to multiple input values.

Example:

for $fruit in ("Apple", "Cherry")
return switch ($fruit)
  case "Apple"
  case "Cherry"
     return "red"
  case "Pear"
     return "green"
  case "Peach"
     return "pink"
  default
     return "unknown"

Result: red red

Expanded QNames

A QName can be prefixed with the letter "Q" and a namespace URI in the Clark Notation.

Examples:

  • Q{http://www.w3.org/2005/xpath-functions/math}pi() returns the number π
  • Q{java:java.io.FileOutputStream}new("output.txt") creates a new Java file output stream

Namespace Constructors

New namespaces can be created via so-called 'Computed Namespace Constructors'.

 
element node { namespace pref { 'http://url.org/' } }

String Concatenations

Two vertical bars || (also named pipe characters) can be used to concatenate strings. This operator is a shortcut for the fn:concat() function.

 
'Hello' || ' ' || 'Universe'

External Variables

Default values can be attached to external variable declarations. This way, an expression can also be evaluated if its external variables have not been bound to a new value.

 
declare variable $user external := "admin";
"User:", $user

Serialization

Serialization parameters can be defined within XQuery expressions. Parameters are placed in the query prolog and need to be specified as option declarations, using the output prefix.

Example:

declare namespace output = "http://www.w3.org/2010/xslt-xquery-serialization";
declare option output:omit-xml-declaration "no";
declare option output:method "xhtml";
<html/>

Result: <?xml version="1.0" encoding="UTF-8"?><html></html>

In BaseX, the output prefix is statically bound and can thus be omitted. Note that all namespaces need to be specified when using external APIs, such as XQJ.

Context Item

The context item can be specified in the prolog of an XQuery expression:

Example:

declare context item := document {
  <xml>
    <text>Hello</text>
    <text>World</text>
  </xml>
};

for $t in .//text()
return string-length($t)

Result: 5 5

Annotations

XQuery 3.0 introduces annotations to declare properties associated with functions and variables. For instance, a function may be declared %public, %private, or %updating.

Example:

declare %private function local:max($x1, $x2) {
  if($x1 > $x2) then $x1 else $x2
};

local:max(2, 3)

Functions

The following functions have been added in the XQuery 3.0 Functions and Operators Specification:

fn:analyze-string* fn:available-environment-variables, fn:element-with-id, fn:environment-variable, fn:filter, fn:fold-left, fn:fold-right, fn:for-each, fn:for-each-pair, fn:format-date, fn:format-dateTime, fn:format-integer, fn:format-number, fn:format-time, fn:function-arity, fn:function-lookup, fn:function-name, fn:generate-id, fn:has-children, fn:head, fn:innermost, fn:outermost, fn:parse-xml, fn:parse-xml-fragment, fn:path, fn:serialize, fn:tail, fn:unparsed-text, fn:unparsed-text-available, fn:unparsed-text-lines, fn:uri-collection

New signatures have been added for the following functions:

fn:document-uri, fn:string-join, fn:node-name, fn:round, fn:data

Changelog

Version 8.4
  • Added: %non-deterministic
Version 8.0
  • Added: %basex:inline, %basex:lazy
Version 7.7
Version 7.3
Version 7.2
Version 7.1
Version 7.0