DB2 Version 10.1 for Linux, UNIX, and Windows

pureXML enhancements

Version 10.1 supports the XML data type and enhances the pureXML® feature to make your data processing even more flexible, faster, and more reliable.

New types supported for indexes over XML
You can now create indexes of type DECIMAL and INTEGER over XML data. In situations where your numeric data is of either INTEGER or DECIMAL type, indexes created as DECIMAL and INTEGER values can potentially provide faster query response times.
Functional XML indexes can speed up queries
Starting in DB2® V10.1, you can create functional XML indexes using the fn:upper-case and fn:exists functions. Indexes created using fn:upper-case can speed up case-insensitive searches of XML data. Indexes created using fn:exists can speed up queries that search for specific elements or for the lack of specific elements.
New binary XML format improves performance for certain Java clients
The new binary XML format provides a faster way to transmit and receive XML data between certain Java™ pureXML applications and a DB2 server Version 10.1. For these Java applications, unnecessary XML parsing costs are eliminated, therefore improving performance.
XML casting and truncation errors reduced
Starting in DB2 V10.1, the DB2 database manager handles XML casting in a similar way as for SQL. Situations that used to raise errors now allow applications to continue. Also, to enhance usability, queries that compare XML data of incompatible types return FALSE instead of returning error SQL16061N.
Performance improved for certain XML queries
In DB2 V10.1, the DB2 server has been optimized to improve performance for certain commonly used queries, such as those that use the XMLTABLE function.
FP1: XML data type support added in global variables and compiled SQL functions
Starting with Version 10.1 Fix Pack 1, you can create global variables of XML data type, you can specify the XML data type in parameters and the RETURNS clause of compiled SQL functions, and you can define local XML variables in compiled SQL functions.