The new version 3.0.0 of the tmql4j query suite was released at google code. In this version tmql4j is more flexible and powerful to satisfy every business use case.

As a major modification, the engine architecture and processing model was changed. The new suite contains two different TMQL runtimes, one for each TMQL draft. The drafts are split to avoid ambiguity and conflicts during the querying process. The stack-based processing model is replaced by a more flexible one to enable multi-threaded optimizations.

Each style of the 2008 draft and each part of the topic map modification language ( TMQL-ML ) has been realized in different modules. Because of that, the user can decide which styles and parts of the query language should be supported.

In addition, a new language module was added to enable flexible template definitions, which enables control of the result format of the querying process in the most powerful way. Templates can be used to return results in HTML, XML, JSON or any other format. The results will be embedded automatically by the query processor.

The new version also comes with a couple of other useful features .

• In contrast to the draft, the engine supports filter after each navigation step.
• New axes, like the typed and variants axis, are added to the navigation handler.
• Result processing has been fully re-implemented, which enables the possibility of empty cells and really atomic cell values.
• The engine supports result column aliases, similar to SQL, which can be used to navigate through the result set.
• The engine supports prepared statements to speed up your application.
• A couple of new functions are added to fetch the ontology of the topic map or topics by their characteristic values.
• The new language version comes with a new group-by-clause, that enables an index-based arrangement of the result set.
• Similar to the select-style, the flwr-style supports the limit-, offset- and unique-clauses.