DataDirect has announced the release of DataDirect XQuery 2.0, an "implementation of XQuery that can query XML, relational data, SOAP messages, EDI, or a combination of data sources." Changes include:

• Configurable performance optimizations and capability to read large documents (gigabytes in size) with the streaming API for XML
• Support of a more recent XQuery specification
• Updated SQL Server 2005 and Oracle 10gR2 support
• Access to EDI, CSV, dBase, and other flat file formats
• Improved support for Java extensions

DataDirect XQuery basically presents an analogue to the JDBC mechanism for querying data. A programmer creates a DDXQDataSource sets any parameters associated with the datasource (i.e., file references or JDBC urls, which specify an XQuery driver), and then creates a statement with an XQuery expression. One thing to note about the JDBC support is that the relational databases supported include versions of DB2, Oracle, SQL Server, and Sybase. Other databases aren't mentioned in documentation; MySQL, PostGreSQL, Derby, HSQLDB, McKoi, and other such open source/free databases may be out of luck, unless coders are willing to translate a resultset into XML for XQuery. A free 15-day trial is available for download.

Generally is this a very interesting theme with magnificent capabilities. The question is: How is the performance ? Native XML-Databases (e.g. the upcoming DB2-Release/Viper) provides properly Support to XQuery in an optimized form, which is executed internally on the database server. Roland SOA Kompetenznetzwerk

actually when xpath and xslt appear , we think it may make a shift on how programmers think, but slow adoption due to late compatible implementation and bad performance made it nearly unused SO as of xquery1.0 and XPath2.0 and the new data model,does any one see a chance to change our thinking and architecturing model to fully adapt these standards for data manipulation OR it will still used marginally for convinience what i really mean is a Xpath2 &XQuery CENTRIC solutions and framework? Do you really intend to accept such approach or what?

XPath: powerful, compact expression language for XML data XQuery: XPath++? Sure, why not. XSLT: YUCK. This needs to be dumped for a language with better syntax. Even LISP is more readable and tolerable than this. Most XSLT ends up tied to java anyway with external calls to java to do things that either XSLT can't do, or doesn't do easily. Question: (I haven't actually looked at how this does XQuery against a RDBMS) Since XML docs are fundamentally hierarchically-organized data, and therefore XPath/XQuery designed to descend hierarchical data structures, isn't there a mismatch applying these languages to graph-based datastructures in relational data models?