Fred, > This shouldn't be too difficult to change. > Classes such as Expression and > Select are responsible for building the results.
Great... I had a [very] quick look; I guess the 2nd half of boolean test() in Expression would need an additional if ( o/o2 == null ) return false or so, and/or some enhancements to Column.compare()... right? For somebody who knows his way around this is probably a 1/2h job! Can anybody on hsqldb-developers spare 1/2h to just get this done, and help Hsqldb move a bit more towards full compliance with SQL standards?! Email me directly if details required, I am willing to test and spend time. > if the purpose is a proper implementation of JDO Well, hsqldb may not become a JDO implemention as such, that's a different business, but I believe hsqldb is very well suited to be a back-end datastore for JDO implementations, stuff that implements the JDO API and writes to JDBC, both commercial as well as interesting looking open one coming up on Jakarta. An actual implemention of the JDO API in hsqldb would be a duplication of efforts and not a natural fit I believe. > first step please submit a full list of > changes that are required for JDO, To make Hsqldb usable as datastore in most common JDO implementations, without much extra work, at this point in time may not actually require much work on the hsqldb side!!! The above "Including null rows in select" is a blocker however. We are hoping to use this as part of an (inhouse) framework and have over 10'000 lines of JUnit test code, of which most already pass with using Hsqldb instead Oracle as a datastore; and other than the above have not identified any other major problem to date. Regards, Michael -----Original Message----- From: Fred Toussi [mailto:[EMAIL PROTECTED]] Sent: Dienstag, 17. September 2002 12:35 To: hsqldb-developers; Michael Vorburger, Vertical*i Subject: Re: [Hsqldb-user] Including null rows in select - bug, option or correct behaviour? This shouldn't be too difficult to change. Classes such as Expression and Select are responsible for building the results. We will only get involved if the purpose is a proper implementation of JDO. If you want to help, as a first step please submit a full list of changes that are required for JDO, including public interfaces and internal modifications. We can move from there. Fred Toussi PS: Please note the list address: [EMAIL PROTECTED] -----Original Message----- Michael Vorburger wrote: Any chance you could point me in the right direction where in the source code we could "patch" this ourselves? PS: As quick background, we are evaluting the use of hsqldb as embedded datastore for a JDO-based application. According to the JDO specification, comparison with null attributes should exclude elements from a result set. Of course you could argue "that's not our problem", and strictly technically speaking I of course agree, however if there is a way around this, it would be great, and I believe could give hsqldb more visibility in the JDO community. I am willing to put in some time to help here... ------------------------------------------------------- This SF.NET email is sponsored by: AMD - Your access to the experts on Hammer Technology! Open Source & Linux Developers, register now for the AMD Developer Symposium. Code: EX8664 http://www.developwithamd.com/developerlab _______________________________________________ hsqldb-developers mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/hsqldb-developers
