JDBC 3.0 compliant applications and applets may connect to a JDBC data source using JDBC javax.sql.DataSource instances. The Virtuoso JDBC 3.0 driver provides an implementation of the javax.sql.DataSource interface in the virtuoso.jdbc3.VirtuosoDataSource class, supporting the following properties:

The Virtuoso JDBC 3.0 driver supports connection pooling.

The virtuoso.jdbc3.VirtuosoDataSource implements the javax.sql.ConnectionPoolDataSource interface. In order to use the connection pooling the administrator must deploy one instance of the virtuoso.jdbc3.VirtuosoDriver in the JNDI repository and set all of it's properties except dataSourceName . This is the "main" connection pooling data source. Then the administrator should deploy a second instance of the virtuoso.jdbc3.VirtuosoDataSource class and set only it's dataSourceName property.

Applications will use the second virtuoso.jdbc3.VirtuosoDataSource instance to get a connection. It will in turn call the first one to obtain all connect info and return the java.sql.Connection instance.

Virtuoso supports the industry standard XA specification for distributed transaction processing. The XA specification defines an interface between the transaction manager (TM) and resource manager (RM) in a distributed transaction system. This is a generic interface and it does not directly address the use of distributed transactions from Java. The Java mapping of the XA interface is defined in Sun Microsystems Java Transaction API (JTA) and JDBC 3.0 specifications. The Virtuoso JDBC 3.0 driver supports the JTA architecture by providing the implementation of JTA resource manager interfaces.

The Virtuoso JDBC 3.0 driver provides the virtuoso.java3.VirtuosoXid , virtuoso.java3.VirtuosoXADataSource , virtuoso.java3.VirtuosoXAConnection , and virtuoso.java3.VirtuosoXAResource classes which implement the interfaces javax.transaction.xa.Xid , javax.transaction.xa.XADataSource , javax.sql.XAConnection , and javax.sql.XAResource respectively. The use if these interfaces is usually transparent for applications and the application developer shouldn't bother with them. They are used only by the JTS transaction manager which normally runs as a part of the J2EE server.

The task of the J2EE server administrator is to setup the necessary Virtuoso XA datasources. The exact procedure of this depends on the J2EE server in use (such as BEA WebLogic, IBM WebSphere, etc). Generally, this includes two steps:

The virtuoso.java3.VirtuosoXADataSource class is derived from virtuoso.java3.VirtuosoDataSource and inherits all of its properties. These properties has to be set as described in the section Virtuoso Driver For JDBC 3.0 javax.sql.DataSource.

For example, the following has to be done in case of Sun's J2EE Reference Implementation.

The Virtuoso JDBC 3.0 driver has two implementations of the javax.sql.RowSet interface - virtuoso.javax.OPLCachedRowSet and virtuoso.javax.OPLJdbcRowSet .

The virtuoso.javax.OPLCachedRowSet class implements a totally disconnected, memory cached rowset and the virtuoso.javax.OPLJdbcRowset class spans the rest of the JDBC API to implement it's methods.

The IRIs and RDF literals, kept in the Virtuoso RDF store are represented by a strings and structures. Thus accessing RDF objects needs special datatypes in order to distinguish strings from IRIs and to get language and datatype of the RDF literals.

Therefore Virtuoso JDBC driver provides following classes for accessing RDF store: virtuoso.jdbc3.VirtuosoExtendedString for IRIs and virtuoso.jdbc3.VirtuosoRdfBox for RDF literal objects.

The class virtuoso.jdbc3.VirtuosoExtendedString will be returned when a string representing an IRI is returned to the client. It has two members "str" and "iriType", the "str" member keeps string representation of the IRI, "iriType" denote regular IRI or blank node with enum values VirtuosoExtendedString.IRI and VirtuosoExtendedString.BNODE.

If the RDF literal object have language or datatype specified then virtuoso.jdbc3.VirtuosoRdfBox will be returned. The following methods could be used :

	  String toString () : returns string representation of the literal
	  String getType () : returns string containing the datatype  of the literal
	  String getLang () : returns language code for the literal
      

The following code snippet demonstrates how to use extension datatypes for RDF

    ... initialization etc. skipped for brevity
    boolean more = stmt.execute("sparql select * from <gr> where { ?x ?y ?z }");
    ResultSetMetaData data = stmt.getResultSet().getMetaData();
    while(more)
    {
	rs = stmt.getResultSet();
	while(rs.next())
	{
	    for(int i = 1;i <= data.getColumnCount();i++)
	    {
		String s = rs.getString(i);
		Object o = ((VirtuosoResultSet)rs).getObject(i);
		if (o instanceof VirtuosoExtendedString) // String representing an IRI
		  {
		    VirtuosoExtendedString vs = (VirtuosoExtendedString) o;
                    if (vs.iriType == VirtuosoExtendedString.IRI && (vs.strType & 0x01) == 0x01)
			System.out.println ("<" + vs.str +">");
		    else if (vs.iriType == VirtuosoExtendedString.BNODE)
			System.out.println ("<" + vs.str +">");
		  }
		else if (o instanceof VirtuosoRdfBox) // Typed literal
		  {
		    VirtuosoRdfBox rb = (VirtuosoRdfBox) o;
		    System.out.println (rb.rb_box + " lang=" + rb.getLang() + " type=" + rb.getType());

		  }
		else if(stmt.getResultSet().wasNull())
		    System.out.println("NULL");
		else //
		{
		    System.out.println(s);
		}

	    }
	}
	more = stmt.getMoreResults();
    }
    ...