<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
                    
  <meta http-equiv="Content-Type"
 content="text/html; charset=iso-8859-1">
                    
  <meta name="GENERATOR"
 content="Mozilla/4.79 [en] (Windows NT 5.0; U) [Netscape]">
<!--
Copyright 2003-2006 Sun Microsystems, Inc.  All Rights Reserved.
DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.

This code is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License version 2 only, as
published by the Free Software Foundation.  Sun designates this
particular file as subject to the "Classpath" exception as provided
by Sun in the LICENSE file that accompanied this code.

This code is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
version 2 for more details (a copy is included in the LICENSE file that
accompanied this code).

You should have received a copy of the GNU General Public License version
2 along with this work; if not, write to the Free Software Foundation,
Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.

Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
CA 95054 USA or visit www.sun.com if you need additional information or
have any questions.
-->
  <title>javax.sql.rowset.spi</title>
   
</head>
  <body bgcolor="#ffffff">

The standard classes and interfaces that a third party vendor has to
use in its implementation of a synchronization provider. These classes and 
interfaces are referred to as the Service Provider Interface (SPI). A vendor may 
have its implementation included on the JDBC web page that lists available
<code>SyncProvider</code> implementations by sending email to <code>jdbc@sun.com</code>.
Doing this helps make developers aware of the implementation. To make it possible
for a <code>RowSet</code> object to use an implementation, the vendor must register
it with the <code>SyncFactory</code> singleton. (See the class comment for
<code>SyncProvider</code> for a full explanation of the registration process and 
the naming convention to be used.)
<P>
<h2>Table of Contents</h2>
<ul>
<li><a href="#pkgspec">1.0 Package Specification</a>
<li><a href="#arch">2.0 Service Provider Architecture</a>
<li><a href="#impl">3.0 Implementer's Guide</a>
<li><a href="#resolving">4.0 Resolving Synchronization Conflicts</a>
<li><a href="#relspec">5.0 Related Specifications</a>
<li><a href="#reldocs">6.0 Related Documentation</a>
</ul>

<h3><a name="pkgspec">1.0 Package Specification</h3>
<P>
The following classes and interfaces make up the <code>javax.sql.rowset.spi</code>
package:
<UL>
  <LI><code>SyncFactory</code>
  <LI><code>SyncProvider</code>
  <LI><code>SyncFactoryException</code>
  <LI><code>SyncProviderException</code>
  <LI><code>SyncResolver</code>
  <LI><code>XmlReader</code>
  <LI><code>XmlWriter</code>
  <LI><code>TransactionalWriter</code>
</UL>
The following interfaces, in the <code>javax.sql</code> package, are also part of the SPI:
<UL>
  <LI><code>RowSetReader</code>
  <LI><code>RowSetWriter</code>
</UL>
<P>
A <code>SyncProvider</code> implementation provides a disconnected <code>RowSet</code>
object with the mechanisms for reading data into it and for writing data that has been
modified in it
back to the underlying data source.  A <i>reader</i>, a <code>RowSetReader</code> or
<code>XMLReader</code> object, reads data into a <code>RowSet</code> object when the 
<code>CachedRowSet</code> methods <code>execute</code> or <code>populate</code> 
are called.  A <i>writer</i>, a <code>RowSetWriter</code> or <code>XMLWriter</code>
object, writes changes back to the underlying data source when the 
<code>CachedRowSet</code> method <code>acceptChanges</code> is called.
<P>
The process of writing changes in a <code>RowSet</code> object to its data source
is known as <i>synchronization</i>.  The <code>SyncProvider</code> implementation that a
<code>RowSet</code> object is using determines the level of synchronization that the
<code>RowSet</code> object's writer uses. The various levels of synchronization are
referred to as <i>grades</i>.
<P>
The lower grades of synchronization are
known as <i>optimistic</i> concurrency levels because they optimistically
assume that there will be no conflicts or very few conflicts.  A conflict exists when
the same data modified in the <code>RowSet</code> object has also been modified 
in the data source. Using the optimistic concurrency model means that if there
is a conflict, modifications to either the data source or the <code>RowSet</code> 
object will be lost.
<P>
Higher grades of synchronization are called <i>pessimistic</i> because they assume
that others will be accessing the data source and making modifications.  These
grades set varying levels of locks to increase the chances that no conflicts
occur.
<P>
The lowest level of synchronization is simply writing any changes made to the 
<code>RowSet</code> object to its underlying data source.  The writer does
nothing to check for conflicts. 
If there is a conflict and the data
source values are overwritten, the changes other parties have made by to the data
source are lost. 
<P>
The <code>RIXMLProvider</code> implementation uses the lowest level 
of synchronization and just writes <code>RowSet</code> changes to the data source.
This is true because  typically XML data sources do not enable transaction
techniques for maintaining the integrity of data. However, specific standards 
groups have considered offering XML-based synchronization.  For details, see
<PRE>
     <a href="http://www.syncml.org">http://www.syncml.org</a>
</PRE>
<P>
For the the next level up, the
writer checks to see if there are any conflicts, and if there are, 
it does not write anything to the data source.  The problem with this concurrency
level is that if another party has modified the corresponding data in the data source 
since the <code>RowSet</code> object got its data,
the changes made to the <code>RowSet</code> object are lost. The
<code>RIOptimisticProvider</code> implementation uses this level of synchronization.
<P>
At higher levels of synchronization, referred to as pessimistic concurrency,
the writer take steps to avoid conflicts by setting locks. Setting locks
can vary from setting a lock on a single row to setting a lock on a table 
or the entire data source. The level of synchronization is therefore a tradeoff 
between the ability of users to access the data source concurrently and the  ability
of the writer to keep the data in the <code>RowSet</code> object and its data source
synchronized.
<P>
It is a requirement that all disconnected <code>RowSet</code> objects 
(<code>CachedRowSet</code>, <code>FilteredRowSet</code>, <code>JoinRowSet</code>, 
and <code>WebRowSet</code> objects) obtain their <code>SyncProvider</code> objects
from the <code>SyncFactory</code> mechanism.  
<P>
The reference implementation (RI) provides two synchronization providers.
	<UL>
       <LI><b><tt>RIOptimisticProvider</tt></b> <br>
           The default provider that the <code>SyncFactory</code> instance will
           supply to a disconnected <code>RowSet</code> object when no provider
           implementation is specified.<BR>
           This synchronization provider uses an optimistic concurrency model,
           assuming that there will be few conflicts among users 
           who are accessing the same data in a database.  It avoids
           using locks; rather, it checks to see if there is a conflict
           before trying to synchronize the <code>RowSet</code> object and the
           data source. If there is a conflict, it does nothing, meaning that
		   changes to the <code>RowSet</code> object are not persisted to the data
           source.
       <LI><B><tt>RIXMLProvider</tt></B> <BR>
            A synchronization provider that can be used with a
            <code>WebRowSet</code> object, which is a rowset that can be written 
            in XML format or read from XML format. The 
            <code>RIXMLProvider</code> implementation does no checking at all for
            conflicts and simply writes any updated data in the
            <code>WebRowSet</code> object to the underlying data source.
            <code>WebRowSet</code> objects use this provider when they are 
            dealing with XML data.
	</UL>

These <code>SyncProvider</code> implementations
are bundled with the reference implementation, which makes them always available to
<code>RowSet</code> implementations. 
<code>SyncProvider</code> implementations make themselves available by being
registered with the <code>SyncFactory</code> singleton.  When a <code>RowSet</code> 
object requests a provider, by specifying it in the constructor or as an argument to the
<code>CachedRowSet</code> method <code>setSyncProvider</code>, 
the <code>SyncFactory</code> singleton
checks to see if the requested provider has been registered with it.
If it has, the <code>SyncFactory</code> creates an instance of it and passes it to the
requesting <code>RowSet</code> object.  
If the <code>SyncProvider</code> implementation that is specified has not been registered,
the <code>SyncFactory</code> singleton causes a <code>SyncFactoryException</code> object
to be thrown.  If no provider is specified,
the <code>SyncFactory</code> singleton will create an instance of the default
provider implementation, <code>RIOptimisticProvider</code>,
and pass it to the requesting <code>RowSet</code> object.

<P>
If a <code>WebRowSet</code> object does not specify a provider in its constructor, the
<code>SyncFactory</code> will give it an instance of <code>RIOptimisticProvider</code>.
However, the constructor for <code>WebRowSet</code> is implemented to set the provider 
to the <code>RIXMLProvider</code>, which reads and writes a <code>RowSet</code> object
in XML format.
<P>
See the <a href="SyncProvider.html">SyncProvider</a> class
specification for further details.
<p>
Vendors may develop a <tt>SyncProvider</tt> implementation with any one of the possible
levels of synchronization, thus giving <code>RowSet</code> objects a choice of
synchronization mechanisms.  A vendor can make its implementation available by 
registering the fully qualified class name with Sun Microsystems at 
<code>jdbc@sun.com</code>. This process is discussed in further detail below. 
<P>

<a name="arch"><h3>2.0 Service Provider Interface Architecture</h3>
<ul>
<b>2.1 Overview</b>
<p>
The Service Provider Interface provides a pluggable mechanism by which
<code>SyncProvider</code> implementations can be registered and then generated when
required. The lazy reference mechanism employed by the <code>SyncFactory</code> limits
unnecessary resource consumption by not creating an instance until it is 
required by a disconnected
<code>RowSet</code> object. The <code>SyncFactory</code> class also provides
a standard API to configure logging options and streams that <b>may</b> be provided
by a particular <code>SyncProvider</code> implementation.
<p>
<b>2.2 Registering with the <code>SyncFactory</code></b>
<p>
A third party <code>SyncProvider</code> implementation must be registered with the 
<code>SyncFactory</code> in order for a disconnected <code>RowSet</code> object 
to obtain it and thereby use its <code>javax.sql.RowSetReader</code> and 
<code>javax.sql.RowSetWriter</code>
implementations. The following registration mechanisms are available to all 
<code>SyncProvider</code> implementations:
<ul>
<li><b>System properties</b> - Properties set at the command line. These
properties are set at run time and apply system-wide per invocation of the Java
application. See the section <a href="#reldocs">"Related Documentation"</a>
further related information.
<p>
<li><b>Property Files</b> - Properties specified in a standard property file.
This can be specified using a System Property or by modifying a standard
property file located in the platform run-time. The
reference implementation of this technology includes a standard property
file than can be edited to add additional <code>SyncProvider</code> objects.
<p> 
<li><b>JNDI Context</b> - Available providers can be registered on a JNDI
context. The <tt>SyncFactory</tt> will attempt to load <tt>SyncProvider</tt>
objects bound to the context and register them with the factory. This
context must be supplied to the <code>SyncFactory</code> for the mechanism to 
function correctly.
</ul>
<p>
Details on how to specify the system properties or properties in a property file
and how to configure the JNDI Context are explained in detail in the
<a href="SyncFactory.html"><code>SyncFactory</code></a> class description.
<p>
<b>2.3 SyncFactory Provider Instance Generation Policies</b>
<p>
The <code>SyncFactory</code> generates a requested <code>SyncProvider</code>
object if the provider has been correctly registered.  The
following policies are adhered to when either a disconnected <code>RowSet</code> object
is instantiated with a specified <code>SyncProvider</code> implementation or is 
reconfigured at runtime with an alternative <code>SyncProvider</code> object.
<ul>
<li> If a <code>SyncProvider</code> object is specified and the <code>SyncFactory</code>
contains <i>no</i> reference to the provider, a <code>SyncFactoryException</code> is 
thrown.
<p>
<li> If a <code>SyncProvider</code> object is specified and the <code>SyncFactory</code>
contains a reference to the provider, the requested provider is supplied.
<p>
<li> If no <code>SyncProvider</code> object is specified, the reference
implementation provider <code>RIOptimisticProvider</code> is supplied.
</ul>
<p>
These policies are explored in more detail in the <a href="SyncFactory.html">
<code>SyncFactory</code></a> class.
</ul>

<li><a name="impl"><h3>3.0 SyncProvider Implementer's Guide</h3>
<ul>
<b>3.1 Requirements</b>
<p>
A compliant <code>SyncProvider</code> implementation that is fully pluggable
into the <code>SyncFactory</code> <b>must</b> extend and implement all
abstract methods in the <a href="SyncProvider.html"><code>SyncProvider</code></a>
class. In addition, an implementation <b>must</b> determine the 
grade, locking and updatable view capabilities defined in the
<code>SyncProvider</code> class definition. One or more of the
<code>SyncProvider</code> description criteria <b>must</b> be supported. It
is expected that vendor implementations will offer a range of grade, locking, and
updatable view capabilities.
<p>
Furthermore, the <code>SyncProvider</code> naming convention <b>must</b> be followed as
detailed in the <a href="SyncProvider.html"><code>SyncProvider</code></a> class
description.
<p>
<b>3.2 Grades</b>
<p>
JSR 114 defines a set of grades to describe the quality of synchronization
a <code>SyncProvider</code> object can offer a disconnected <code>RowSet</code>
object. These grades are listed from the lowest quality of service to the highest.
<ul>
<li><b>GRADE_NONE</b> - No synchronization with the originating data source is
provided. A <code>SyncProvider</code> implementation returning this grade will simply
attempt to write any data that has changed in the <code>RowSet</code> object to the
underlying data source, overwriting whatever is there. No attempt is made to compare 
original values with current values to see if there is a conflict. The 
<code>RIXMLProvider</code> is implemented with this grade.
<p>
<li><b>GRADE_CHECK_MODIFIED_AT_COMMIT</b> - A low grade of optimistic synchronization.
A <code>SyncProvider</code> implementation returning this grade
will check for conflicts in rows that have changed between the last synchronization 
and the current synchronization under way. Any changes in the originating data source
that have been modified will not be reflected in the disconnected <code>RowSet</code>
object. If there are no conflicts, changes in the <code>RowSet</code> object will be
written to the data source. If there are conflicts, no changes are written.
The <code>RIOptimisticProvider</code> implementation uses this grade.
<p>
<li><b>GRADE_CHECK_ALL_AT_COMMIT</b> - A high grade of optimistic synchronization.
A <code>SyncProvider</code> implementation   returning this grade
will check all rows, including rows that have not changed in the disconnected
<code>RowSet</code> object. In this way, any changes to rows in the underlying
data source will be reflected in the disconnected <code>RowSet</code> object
when the synchronization finishes successfully.
<p>
<li><b>GRADE_LOCK_WHEN_MODIFIED</b> - A pessimistic grade of synchronization.
<code>SyncProvider</code> implementations returning this grade will lock
the row in the originating  data source that corresponds to the row being changed
in the <code>RowSet</code> object to reduce the possibility of other
processes modifying the same data in the data source.
<p>
<li><b>GRADE_LOCK_WHEN_LOADED</b> - A higher pessimistic synchronization grade.
A <code>SyncProvider</code> implementation returning this grade will lock
the entire view and/or  table affected by the original query used to
populate a <code>RowSet</code> object.
</ul>
<p>
<b>3.3 Locks</b>
<p>
JSR 114 defines a set of constants that specify whether any locks have been
placed on a <code>RowSet</code> object's underlying data source and, if so,
on which constructs the locks are placed.  These locks will remain on the data
source while the <code>RowSet</code> object is disconnected from the data source.
<P>
These constants <b>should</b> be considered complementary to the 
grade constants. The default setting for the majority of grade settings requires
that no data source locks remain when a <code>RowSet</code> object is disconnected 
from its data source.
The grades <code>GRADE_LOCK_WHEN_MODIFIED</code> and
<code>GRADE_LOCK_WHEN_LOADED</code> allow a disconnected <code>RowSet</code> object
to have a fine-grained control over the degree of locking.
<ul>
<li><b>DATASOURCE_NO_LOCK</b> - No locks remain on the originating data source. 
This is the default lock setting for all <code>SyncProvider</code> implementations 
unless otherwise directed by a <code>RowSet</code> object.
<p>
<li><b>DATASOURCE_ROW_LOCK</b> - A lock is placed on the rows that are touched by
the original SQL query used to populate the <code>RowSet</code> object.
<p> 
<li><b>DATASOURCE_TABLE_LOCK</b> - A lock is placed on all tables that are touched
by the query that was used to populate the <code>RowSet</code> object.
<p>
<li><b>DATASOURCE_DB_LOCK</b>
A lock is placed on the entire data source that is used by the <code>RowSet</code>
object.
</ul>
<p>
<b>3.4 Updatable Views</b>
<p>
A <code>RowSet</code> object may be populated with data from an SQL <code>VIEW</code>.
The following constants indicate whether a <code>SyncProvider</code> object can
update data in the table or tables from which the <code>VIEW</code> was derived.
<ul>
<li><b>UPDATABLE_VIEW_SYNC</b>
Indicates that a <code>SyncProvider</code> implementation  supports synchronization
to the table or tables from which the SQL <code>VIEW</code> used to populate  a
a <code>RowSet</code> object is derived.
<p>
<li><b>NONUPDATABLE_VIEW_SYNC</b>
Indicates that a <code>SyncProvider</code> implementation  does <b>not</b> support
synchronization to the table or tables from which the SQL <code>VIEW</code> 
used to populate  a <code>RowSet</code> object is derived.
</ul>
<p>
<b>3.5 Usage of <code>SyncProvider</code> Grading and Locking</b>
<p>
In the example below, the reference <tt>CachedRowSetImpl</tt> implementation
reconfigures its current <tt>SyncProvider</tt> object by calling the 
<tt>setSyncProvider</tt> method.<br>
       
<PRE>
    CachedRowSetImpl crs = new CachedRowSetImpl();
    crs.setSyncProvider("com.foo.bar.HASyncProvider");
</PRE>
    An application can retrieve the <tt>SyncProvider</tt> object currently in use
by a disconnected <code>RowSet</code> object. It can also retrieve the
grade of synchronization with which the provider was implemented and the degree of
locking currently in use.  In addition, an application has the flexibility to set
the degree of locking to be used, which can increase the possibilities for successful
synchronization.  These operation are shown in the following code fragment.
<PRE>
    SyncProvider sync = crs.getSyncProvider();

    switch (sync.getProviderGrade()) {
    case: SyncProvider.GRADE_CHECK_ALL_AT_COMMIT
         //A high grade of optimistic synchronization
    break;
    case: SyncProvider.GRADE_CHECK_MODIFIED_AT_COMMIT 
         //A low grade of optimistic synchronization 
    break;
    case: SyncProvider.GRADE_LOCK_WHEN_LOADED 
         // A pessimistic synchronization grade 
    break;
    case: SyncProvider.GRADE_LOCK_WHEN_MODIFIED 
         // A pessimistic synchronization grade 
    break;
    case: SyncProvider.GRADE_NONE 
      // No synchronization with the originating data source provided
    break;
    }
	  
    switch (sync.getDataSourcLock() {
      case: SyncProvider.DATASOURCE_DB_LOCK
       // A lock is placed on the entire datasource that is used by the
       // <code>RowSet</code> object 
       break;

      case: SyncProvider.DATASOURCE_NO_LOCK
       // No locks remain on the  originating data source.
      break;

      case: SyncProvider.DATASOURCE_ROW_LOCK
       // A lock is placed on the rows that are  touched by the original 
       // SQL statement used to populate
       // the RowSet object that is using the SyncProvider
       break;

      case: DATASOURCE_TABLE_LOCK
       // A lock is placed on  all tables that are touched by the original 
       // SQL statement used to populated
       // the RowSet object that is using the SyncProvider
       break;

</PRE>
    It is also possible using the static utility method in the
<code>SyncFactory</code> class to determine the list of <code>SyncProvider</code>
implementations currently registered with the <code>SyncFactory</code>.
       
<pre>
	Enumeration e = SyncFactory.getRegisteredProviders();
</pre>    

</ul>

<h3><a name="resolving">4.0 Resolving Synchronization Conflicts</h3>

The interface <code>SyncResolver</code> provides a way for an application to
decide manually what to do when a conflict occurs. When the <code>CachedRowSet</code>
method <code>acceptChanges</code> finishes and has detected one or more conflicts,
it throws a <code>SyncProviderException</code> object.  An application can
catch the exception and
have it retrieve a <code>SyncResolver</code> object by calling the method
<code>SyncProviderException.getSyncResolver()</code>. 
<P>
A <code>SyncResolver</code> object, which is a special kind of 
<code>CachedRowSet</code> object or
a <code>JdbcRowSet</code> object that has implemented the <code>SyncResolver</code> 
interface,  examines the conflicts row by row. It is a duplicate of the 
<code>RowSet</code> object being synchronized except that it contains only the data
from the data source this is causing a conflict. All of the other column values are
set to <code>null</code>. To navigate from one conflict value to another, a
<code>SyncResolver</code> object provides the methods <code>nextConflict</code> and
<code>previousConflict</code>.
<P>
The <code>SyncResolver</code> interface also
provides methods for doing the following:
<UL>
 <LI>finding out whether the conflict involved an update, a delete, or an insert
 <LI>getting the value in the data source that caused the conflict
 <LI>setting the value that should be in the data source if it needs to be changed
     or setting the value that should be in the <code>RowSet</code> object if it needs
     to be changed
</UL>
<P>
When the <code>CachedRowSet</code> method <code>acceptChanges</code> is called, it 
delegates to the <code>RowSet</code> object's  <code>SyncProvider</code> object.
How the writer provided by that <code>SyncProvider</code> object is implemented
determines what level (grade) of checking for conflicts will be done.  After all 
checking for conflicts is completed and one or more conflicts has been found, the method
<code>acceptChanges</code> throws a <code>SyncProviderException</code> object. The
application can catch the exception and use it to obtain a <code>SyncResolver</code> object.  
<P>
The application can then use <code>SyncResolver</code> methods to get information
about each conflict and decide what to do.  If the application logic or the user
decides that a value in the <code>RowSet</code> object should be the one to
persist, the application or user can overwrite the data source value with it.  
<P>
The comment for the <code>SyncResolver</code> interface has more detail.

<a name="relspec"><h3>5.0 Related Specifications</h3>
<ul>
<li><a href="http://java.sun.com/products/jndi">JNDI 1.3</a>
<li><a href="{@docRoot}/../technotes/guides/logging/index.html">Java Logging
APIs</a>
</ul>
<a name="reldocs"><h3>6.0 Related Documentation</h3>
<ul>
<li><a href="{@docRoot}/../technotes/tools/index.html#basic">System
properties</a>
<li>Resource Files
<li><a href="http://java.sun.com/tutorial/jdbc">DataSource for JDBC
Connections</a>
</ul>

</body>
</html>