Changes

Jump to navigation Jump to search
16,158 bytes added ,  16:04, 19 July 2021
This article is part of the [[QueryingXQuery|Query XQuery Portal]].It demonstrates two different ways to invoke Java code from XQuery, and (since {{Version|7.2.1}}) an extension it presents extensions to make Java code aware of access the current query contextfrom Java.
The Java Binding [[Querying|query]] feature is an extensibility mechanism which enables developersto directly access Java variables and execute code from XQuery. Addressed Java code must either be contained in the Java classpath, or it must be located in the [[Repository]]. Please bear in mind that the execution of Java classes code may cause side effects that conflict with the functional nature of XQuery, or may introduce new security risks to your project. {{Mark|Updated with Version 9.6:}}* With the middle dot notation, three adjacent dots can be used to specify array types.* The path to the standard package {{Code|java.lang.}} can now be omitted.* Java objects are now wrapped into function items.* Results of constructor calls are always returned as function item.* A new option {{Option|WRAPJAVA}} was added to control how Java values are converted to XQuery.* The Mapping rules were refined and unified. The most important changes:** {{Code|array(*)}} type added.** {{Code|xs:integer}} values are converted to {{Code|long}} values.** {{Code|xs:unsignedShort}} values are converted to {{Code|char}} values.* All error messages were revised and improved. =Identification= ==Classes== A Java class is identified bya namespace URI. The original URI is rewritten as follows: # The [[#URI Rewriting|URI Rewriting]] steps are applied to the URI.# Slashes in the resulting URI are replaced with dots.namespaces; # The last path segment of the URI is capitalized and rewritten to [https://en.wikipedia.org/wiki/CamelCase CamelCase]. The normalization steps are skipped if the namespace URI must is prefixed with {{Code|java:}}. The path to the standard package {{Code|java.lang.}} can be omitted: * <code><nowiki>http://basex.org/modules/meta-data</nowiki></code> → <code>org.basex.modules.MetaData</code>* <code>java:java.lang.String</code> → <code>java.lang.String</code>* <code>StringBuilder</code> → <code>java.lang.StringBuilder</code> ==Functions and Variables== Java constructors, functions and variables can be referenced and evaluated by the existing XQuery function syntax: * The namespace of the function name identifies the Java class.* The local part of the form name, which is rewritten to camel case, identifies a variable or function of that class.* The middle dot character <code>[https://www.fileformat.info/info/unicode/char/b7/index.htm ·]</code> (<code>&amp;#xB7;</code>, a valid character in XQuery names, but not in Java) can be used to append exact Java parameter types to the function name. Class types must be referenced by their full path. Three adjacent dots can be used to address an array argument. {| class="wikitable"|- valign="top"! Addressed code! XQuery! Java|- valign="top"| Variable| <code>Q{MonoInteger}MIN_VALUE()</code>|java<code>Integer.MIN_VALUE</code>|- valign="top"| Function| <code>Q{Object}hash-code($object)</code>| <code>object.hashCode()</code>|- valign="top"| Function with argument| <code>Q{String}split·String·int($string, ';', xs:int(3))</code>| <code>string.split(";", 3)</code>|- valign="top"| Constructor with array argument| <code>Q{String}new·byte...(xs:fullyhexBinary('414243'))</code>| <code>new String(new byte[] { 41, 42, 43 })</code>|} As XQuery and Java have different type systems, XQuery arguments must be converted to equivalent Java values, and the result of a Java function is converted back to an XQuery value (see [[#Data Types|Data Types]]).qualified If the Java function you want to address is not detected, you may need to cast your values to the target type.ClassNameFor example, if a Java function expects a primitive {{Code|int}} value, you will need to convert your XQuery integers to {{Code|xs:int}}.
=Namespace Declarations=
The In the following example uses Java’s , the Java {{MonoCode|Math}} class and is referenced. When executed, the query returns the cosine of an angle.The by calling the static method {{MonoCode|cos()}} method can be directly called, as it is a and the value of π by addressing the static variable via {{MonoCode|staticPI()}} method<syntaxhighlight lang="xquery">declare namespace math = "java:java.lang.Math";math:cos(xs:double(0)), math:PI()</syntaxhighlight>
<pre class="brush:xquery">declare namespace math = "java:javaWith the [[XQuery 3.lang0#Expanded QNames|Expanded QName]] notation of XQuery 3.Math";math:cos(xs0, the namespace can directly be embedded in the function call:double(0))</pre>
The next example writes 256 bytes to the file <syntaxhighlight lang="xquery">Q{{Mono|outputjava:java.txt}}lang.First, a new {{Mono|FileWriterMath}} instance is createdcos(xs: by calling the{{Mono|newdouble(0))}} function, the class constructor is invoked. Instancemethods are called by passing on the resulting Java object asfirst argument:</syntaxhighlight>
The constructor of a class can be invoked by calling the virtual function {{Code|new()}}. Instance methods can then called by passing on the resulting Java object as first argument. In the following example, 256 bytes are written to the file {{Code|output.txt}}. First, a new {{Code|FileWriter}} instance is created, and its {{Code|write()}} function is called in the next step: <pre classsyntaxhighlight lang="brush:xquery">declare namespace fw = &quot;'java:java.io.FileWriter&quot;'; let $file := fw:new('output.txt')
return (
for $i in 0 to 255
fw:close($file)
)
</presyntaxhighlight>
If the result of a Java call contains invalid XML characters, it will be rejected. The validity check can be disabled by setting {{Option|CHECKSTRINGS}} to false. In generalthe example below, it a file with a single {{Code|00}} byte is recommended to use XQuery expressions written, and this file will then be accessed by via Java functions whenever possible,: <syntaxhighlight lang="xquery">declare namespace br = 'java:java.io.BufferedReader';declare namespace fr = 'java:java.io.FileReader'; declare option db:checkstrings 'false'; (: write file :)as Java code cannot be prefile:write-compiledbinary('00.bin', xs:hexBinary('00')), and will often be evaluated slower than optimizedXQuery code(: read file :)let $br := br:new(fr:new('00. Nextbin'))return ( br:readLine($br), Java code br:close($br))</syntaxhighlight> The option can only also be executed with [[User_Management|admin permissions]]specified via a pragma: <syntaxhighlight lang="xquery">(# db:checkstrings #) { br:new(fr:new('00.bin')) ! (br:readLine(.), br:close(.))}</syntaxhighlight>
=Module Imports=
{{Mark|Introduced with Version 7A Java classes can also be instantiated by ''importing'' them as a module: A new instance of the addressed class will be constructed, which can then be referenced in the query body.2.1:}}
Java code can also be integrated by ''importing'' classes as modules. In this casethe (side-effecting) example below,a new HashSet instance is created, values are added, and the size of the addressed class set is createdreturned. As {{Code|set:add()}} returns boolean values, which can then be accessed in {{Function|Profiling|prof:void}} is used to swallow the query body.values:
An example <syntaxhighlight lang="xquery">import module namespace set = "java:java.util.HashSet";prof:void( for $s in (the boolean values returned by {{Mono|"one", "two", "one") return set:add($s)}} are ignored),set:size()</syntaxhighlight>
<pre The execution of imported classes is more efficient than the execution of instances that have been created via {{Code|new()}}. In turn, no arguments can be supplied in the import statement, and the construction will only be successful if the class="brush:xquery">import module namespace set = "java:java.utilcan be instantiated without arguments.HashSet";let $loop := for $i in 1 to 10000 return set:add($i)return set:size()</pre>
Advantages of this approach are:* imported code can be executed faster than instances created at runtime via {{Mono|new()}}.* the work on class instances ensures that queries run in parallel will not cause any concurrency issues (provided that the class contains no static variables or functions).=Integration=
A drawback is that no arguments Java classes can be passed on coupled more closely to BaseX. If a class inherits the abstract [https://github.com/BaseXdb/basex/blob/master/basex-core/src/main/java/org/basex/query/QueryModule.java QueryModule] class constructor, the two variables [https://github.This also implies that com/BaseXdb/basex/blob/master/basex-core/src/main/java/org/basex/query/QueryContext.java queryContext] and [https://github.com/BaseXdb/basex/blob/master/basex-core/src/main/java/org/basex/query/StaticContext.java staticContext] get available, which provide access to the class must provide global and static context of a constructor without no argumentsquery.
=ContextThe [https://github.com/BaseXdb/basex/blob/master/basex-Awareness=core/src/main/java/org/basex/query/QueryResource.java QueryResource] interface can be implemented to enforce finalizing operations, such as the closing of opened connections or resources in a module. Its {{Code|close()}} method will be called after the XQuery expression has been fully evaluated.
{{Mark|Introduced with Version 7.2.1:}}==Annotations==
If an instantiated class inherits the abstract [https://github.com/BaseXdb/basex/blob/master/src/main/java/org/basex/query/QueryModule.java QueryModule] class The internal properties of BaseX, it will functions can be enriched with two internal, context-dependent variablesassigned via annotations:
* Java functions can only be executed by users with [[User_Management|Admin permissions]]. You can annotate a function with {{MonoCode|context@Requires(<Permission>)}} to also make it accessible to users with fewer privileges.* Java code is an instance of treated as ''non-deterministic'', as its behavior cannot be predicted by the [https://githubXQuery processor.com/BaseXdb/basex/blob/master/src/main/java/org/basex/query/QueryContext.java QueryContext] class. It provides access to all static You may annotate a function as {{Code|@Deterministic}} if you know that it will have no side effects and dynamic properties of will always yield the current querysame result.* Java code is treated as ''context-independent''. If a function accesses the query context, it should be annotated as {{MonoCode|input@ContextDependent}} * Java code is an instance of [https://github.com/BaseXdb/basex/blob/master/src/main/java/org/basex/util/InputInfo.java InputInfo]treated as ''focus-independent''. It contains line/column and If a function accesses the original query and is mainly used for providing better error messages.current context item, position or size, it should be annotated as {{Code|@FocusDependent}}
Context-aware classes represent In the following code, information from the static query context is returned by the first function, and a powerful interface for writing Java modules that are more closely coupled with query exception is raised by the BaseX core library.second function:
The following XQuery code invokes two <syntaxhighlight lang="xquery">import module namespace context-aware Java methods:= 'org.basex.examples.query.ContextModule';
<pre class="brushelement user { context:xquery">user()},try {import module namespace element to-int { context = :to-int('abc'java) }} catch basex:org.basex.examples.query.ContextModule';error { element error { $err:description }}</syntaxhighlight>
<context>{ contextThe imported Java class is shown below:function-namespace()}</context>,<to-int>{ try { context:to-int('abc') } catch * { 'Error in line', $err:line-number }}</to-int></pre>
The context-aware Java class is shown below:<syntaxhighlight lang="java">package org.basex.examples.query;
<pre class="brush:java">
import org.basex.query.*;
import org.basex.query.value.item.*;import org.basex.util.*;
/**
* This example class contains two methods for demonstrating inherits the use of the * {@link QueryModule} classand * implements the QueryResource interface.
*/
public class QueryModuleExample ContextModule extends org.basex.query.QueryModule implements QueryResource {
/**
* Returns the root expression name of this query as stringthe logged-in user. * @return root expressionuser string
*/
@Requires(Permission.NONE) @Deterministic @ContextDependent public String rootExpressionuser() { return queryContext.context.rootuser.toString()name;
}
/**
* Converts the specified string to an integer.
* @param value string representationto be converted * @return resulting integer
* @throws QueryException query exception
*/
@Requires(Permission.NONE)
@Deterministic
public int toInt(final String value) throws QueryException {
try {
return Integer.parseInt(value);
} catch(NumberFormatException ex) {
throw new QueryException(input, new QNm("FORM0001Integer conversion failed: "), ex.getMessage()+ value);
}
}
 
@Override
public void close() {
// defined in QueryResource interface, will be called after query evaluation
}
}
</presyntaxhighlight>
The result will look as follows:
<pre classsyntaxhighlight lang="brush:xml"><contextuser>admin</admin>http<error>Integer conversion failed: abc</error></syntaxhighlight> Please visit the XQuery 3.0 specification if you want to get more insight into[https://www.w3.org/2005TR/xpath-functions-31/#properties-of-functions function properties]. ==Updates== The {{Code|@Updating}} annotation can be applied to mark Java functions that perform write or update operations: <syntaxhighlight lang="java"> @Updating public void backup() { // ... }</syntaxhighlight> An XQuery expression will be handled as an [[XQuery Update#Updating Expressions|updating expression]] if it calls an updating Java function. In contrast to XQuery update operations, the Java code will immediately be executed, but the result will be cached as if {{Function|Update|update:output}} was called. The annotation is particularly helpful if combined with a lock annotation. ==Locking== By default, a Java function will be executed in parallel with other code. If a Java function performs sensitive operations, it is advisable to explicitly lock the code. ===Java Locks=== Java provides a handful of mechanism to control the execution of code. The concurrent execution of functions can be avoided with the {{Code|synchronized}} keyword. For more complex scenarios, the Lock, Semaphore and Atomic classes can be brought into play. ===XQuery Locks=== If you want to synchronize the execution of your code with BaseX locks, you can take advantage of the {{Code|@Lock}} annotation: <syntaxhighlight lang="java"> @Lock("HEAVYIO") public void read() { // ... }  @Updating @Lock("HEAVYIO") public void write() { // ... }</syntaxhighlight> If an XQuery expression invokes {{Code|write()}}, any other query that calls {{Code|write()}} or {{Code|read()}} needs to wait for the query to be finished. The {{Code|read()}} function can be run in parallel; whereas queries will be queued if {{Code|write()}} is called. More details on concurrent querying can be found in the article on [[Transaction Management]]. ==Data Types== ===Conversion to Java=== Before Java code is executed, the arguments are converted to Java values, depending on the addressed function or constructor parameters. The accepted Java types and the original XQuery types are depicted in the second and first column of the table below. ===Conversion to XQuery=== By default, Java values with the most common types (as shown in the second and third column of the table) are converted to XQuery values. All other values are returned as ''Java items'', which are function items with a wrapped Java value. The results of constructor calls are always returned as Java items. The conversion of the wrapped Java value to XQuery is enforced by invoking the function item: Values in {{Code|Iterator}} and {{Code|Iterable}} instances (Lists, Sets and Collections) are converted to items, and maps are converted to XQuery maps: <syntaxhighlight lang="xquery">declare namespace Scanner = 'java:java.util.Scanner';let $scanner := Scanner:new("A B C") => Scanner:useDelimiter(" ")return $scanner()</contextsyntaxhighlightIf no conversion is defined, a string is returned, resulting from the {{Code|toString()}} method of the object. This method is also called is the string representation of a Java item is requested: <syntaxhighlight lang="xquery">(: returns the string representations of a HashMap and an ArrayList instance :)'Map: ' || Q{java.util.HashMap}new(),string(Q{java:java.util.ArrayList}new())</syntaxhighlight> The conversion can be further controlled with the {{Option|WRAPJAVA}} option. The following values exist: {| class="wikitable"|- valign="top"! Value! Description|- valign="top"| {{Code|some}}| The default: Java values of the most common types are converted, others are wrapped into Java items.|- valign="top"| {{Code|none}}| All Java values are converted. If no conversion is defined, a string is returned, resulting from the {{Code|toString()}} method.|- valign="top"| {{Code|all}}| Java values are wrapped into Java items (excluding those inheriting the internal type {{Code|org.basex.query.value.Value}}).|- valign="top"| {{Code|instance}}| If the method of a class instance was called, the Java value is ignored and the instance is wrapped into a Java item. Otherwise, the Java value is returned.|- valign="top"| {{Code|void}}| Java values are ignored, and an empty sequence is returned instead.|} In the following example, the result of the first function – a char array – is wrapped and passed on toa {{Code|CharBuffer}} function. Without the option, the single-value array would be converted to an {{Code|xs:unsignedShort}} item and the second function call would fail: <syntaxhighlight lang="xquery">(: Without the pragma, the result of toChars would be converted to an xs:unsignedShort item, and the second function call would fail :) (# db:wrapjava all #) { Q{Character}toChars(xs:int(33)) => Q{java.nio.CharBuffer}wrap()}</syntaxhighlight> The next example demonstrates a use case for the {{Code|instance}} option: <syntaxhighlight lang="xquery">(: Thanks to the pragma, the function calls can be chained :) declare namespace set = 'java:java.util.HashSet';let $set := (# db:wrapjava instance #) { set:new() =>Error set:add('1') => set:add('2')}return $set()</syntaxhighlight> The {{Code|void}} option is helpful if side-effecting methods return values that do not contribute to the final result: <syntaxhighlight lang="xquery">(: Without the pragma, 100 booleans would be returned by the FLWOR expression :) declare namespace set = 'java:java.util.HashSet';let $set := set:new()return ( (# db:wrapjava void #) { for $i in line 61 to 100 return set:add($set, $i) }, $set())</tosyntaxhighlight> The irrelevant results could also be swallowed with {{Function|Profiling|prof:void}}. {| class="wikitable"|- valign="top"! XQuery input! Expected or returned Java type! XQuery output|- valign="top"| <code>item()*</code> (no conversion)| <code>org.basex.query.value.Value</code>| <code>item()*</code> (no conversion)|- valign="top"| <code>empty-sequence()</code>| <code>null</code>| <code>empty-sequence()</code>|- valign="top"| <code>xs:string</code>, <code>xs:untypedAtomic</code>| <code>String</code>| <code>xs:string</code>|- valign="top"| <code>xs:unsignedShort</code>| <code>char</code>, <code>Character</code>| <code>xs:unsignedShort</code>|- valign="top"| <code>xs:boolean</code>| <code>boolean</code>, <code>Boolean</code>| <code>xs:boolean</code>|- valign="top"| <code>xs:byte</code>| <code>byte</code>, <code>Byte</code>| <code>xs:byte</code>|- valign="top"| <code>xs:short</code>| <code>short</code>, <code>Short</code>| <code>xs:short</code>|- valign="top"| <code>xs:int</code>| <code>int</code>, <code>Integer</code>| <code>xs:int</code>|- valign="top"| <code>xs:integer</code>, <code>xs:long</code>| <code>long</code>, <code>Long</code>| <code>xs:integer</code>|- valign="top"| <code>xs:unsignedLong</code>| <code>java.math.BigInteger</code>| <code>xs:unsignedLong</code> (lossy)|- valign="top"| <code>xs:decimal</code>| <code>java.math.BigDecimal</code>| <code>xs:decimal</code>|- valign="top"| <code>xs:float</code>| <code>float</code>, <code>Float</code>| <code>xs:float</code>|- valign="top"| <code>xs:double</code>| <code>double</code>, <code>Double</code>| <code>xs:double</code>|- valign="top"| <code>xs:QName</code>| <code>javax.xml.namespace.QName</code>| <code>xs:QName</code>|- valign="top"| <code>xs:anyURI</code>| <code>java.net.URI</code>, <code>java.net.URL</code>| <code>xs:anyURI</code>|- valign="top"| <code>xs:date</code>| <code>javax.xml.datatype.XMLGregorianCalendar</code>| <code>xs:date</code>|- valign="top"| <code>xs:duration</code>| <code>javax.xml.datatype.Duration</code>| <code>xs:duration</code>|- valign="top"| <code>node()</code>| <code>org.w3c.dom.Node</code>| <code>node()</code>|- valign="top"| <code>array(xs:boolean)</code>| <code>boolean[]</code>| <code>xs:boolean*</code>|- valign="top"| <code>array(xs:string)</code>| <code>String[]</code>| <code>xs:string*</code>|- valign="top"| <code>array(xs:unsignedShort)</code>| <code>char[]</code>| <code>xs:unsignedShort*</code>|- valign="top"| <code>array(xs:short)</code>| <code>short[]</code>| <code>xs:short*</code>|-valign="top"| <code>array(xs:int)</code>| <code>int[]</code>| <code>xs:int*</precode>|- valign="top"| <code>array(xs:integer)</code>, <code>array(xs:long)</code>| <code>long[]</code>| <code>xs:integer*</code>|- valign="top"| <code>array(xs:float)</code>| <code>float[]</code>| <code>xs:float*</code>|- valign="top"| <code>array(xs:double)</code>| <code>double[]</code>| <code>xs:double*</code>|- valign="top"| <code>Object[]</code> (others)| <code>item()*</code>| <code>array(*)</code> (others)|- valign="top"| <code>map(*)</code>| java.util.HashMap| <code>Wrapped Java object</code>|} ==URI Rewriting== Before a Java class or module is accessed, its namespace URI will be normalized: # If the URI is a URL:## colons will be replaced with slashes,## in the URI authority, the order of all substrings separated by dots is reversed, and## dots in the authority and the path are replaced by slashes. If no path exists, a single slash is appended.# Otherwise, if the URI is a URN, colons will be replaced with slashes.# Characters other than letters, dots and slashes will be replaced with dashes.# If the resulting string ends with a slash, the {{Code|index}} string is appended. If the resulting path has no file suffix, it may point to either an XQuery module or a Java archive: * {{Code|<nowiki>http://basex.org/modules/hello/World</nowiki>}} → {{Code|org/basex/modules/hello/World}}* {{Code|<nowiki>http://www.example.com</nowiki>}} → {{Code|com/example/www/index}}* {{Code|a/little/example}} → {{Code|a/little/example}}* {{Code|a:b:c}} → {{Code|a/b/c}}
=Changelog=
===; Version 79.26* Updated: {{Anchor|Data Types}}: {{Code|array(*)}} type added; {{Code|xs:integer}} are converted to {{Code|long}} values.1=== ; Version 9.4* Added: Annotation for [[#Updates|updating functions]].* Updated: Single annotation for read and write locks. ; Version 8.4* Updated: Rewriting rules
;Version 8.2
* Added: [[#URI Rewriting|URI Rewriting]]: support for URNs
 
; Version 8.0
* Added: {{Code|QueryResource}} interface, called after a query has been fully evaluated.
 
; Version 7.8
* Added: Java locking annotations
* Updated: {{Code|context}} variable has been split into {{Code|queryContext}} and {{Code|staticContext}}.
 
; Version 7.2.1
* Added: import of Java modules, context awareness
 * Added: [[Category:XQuery#Packaging|Packaging]], [[Category:API#URI Rewriting|URI Rewriting]]
Bureaucrats, editor, reviewer, Administrators
13,550

edits

Navigation menu