JDO 2 has overhauled the query support. JDO-QL has been greatly enhanced (projections, aggregates, single string format, and more), and native SQL support is available.

TheServerSide is trying to get more information out of the JCP via series of articles like this series. The program, TheServerSide Community Process is there to allow spec leads, and expert groups to get the word out earlier.

Read JDO Community Process: JDO 2 Queries - Part 1

For anyone interested in playing with JDO 2 query features, Kodo JDO supports almost all of them today. This includes the features in the upcoming Parts II, III, and IV of the JDOQL series, for those of you who can't wait. We have ready-made sample models and a graphical query browser (pictured in the article) to get you up and experimenting with JDOQL quickly.

native SQL support is available.

Thanks, it is very usefull.

Versant Open Access (formerly JDO Genie) also supports nearly all of JDO 2 JDOQL. Our GUI Workbench includes an interactive query builder. You can develop your query, edit the objects that come back, generate and edit corresponding Java code and execute it on the fly using BeanShell.

David Tinker, Versant

Thanks for much needed explanation of JDO 2 queries. The example SQL show database attributes such as JDOID, JDOCLASS, JDOVERSION.

What works when the database was designed by a DBA and it has no JDO specific fields? I can't tell a company they can use JDO with their legacy database if they must add JDO specific columns to it.

Thanks for much needed explanation of JDO 2 queries. The example SQL show database attributes such as JDOID, JDOCLASS, JDOVERSION.What works when the database was designed by a DBA and it has no JDO specific fields? I can't tell a company they can use JDO with their legacy database if they must add JDO specific columns to it.

this is related to the Kodo implementation and not JDO, but I think the code in his example is based on Kodo defaults. You will probably have a way to configure/remove these additional columns.

Thanks for much needed explanation of JDO 2 queries. The example SQL show database attributes such as JDOID, JDOCLASS, JDOVERSION.What works when the database was designed by a DBA and it has no JDO specific fields? I can't tell a company they can use JDO with their legacy database if they must add JDO specific columns to it.

These are all Kodo defaults, but certainly aren't required. In fact, I think you'd find that Kodo has some of the most flexible ORM around. Though Kodo is happy to auto-generate a schema, most of our customers have existing data and use Kodo to map to existing schemas.

Robin showed an instanceOf query, how would you do that without JDOCLASS attributes?

Robin showed an instanceOf query, how would you do that without JDOCLASS attributes?

It depends on how the class is mapped. Typically when you use inheritance there is going to be some column or combination of columns that are used to differentiate a record of one type from a record of another. The column names don't have to be JDOCLASS and the values don't have to be class names, but the strategy of checking these columns for the proper value is the same.

In other cases where there are no differentiating columns, you might have to outer join to the table for the subclass being "instanceof" tested and make sure a corresponding record exists. In still other cases, it might be impossible to do an instanceof test efficiently at all. Luckily, those cases are rare.