Hi Marilene
Comments inline
On 4/13/07, Marilene Noronha Roder <[EMAIL PROTECTED]> wrote:
Hi Luciano
I was looking at the "old" proposed list of key feautures to be included
in the new DAS M3 release, and I have a question about this one:
Optimisc concurrency control
- Support for OCC recovery (TUSCANY-916) - Default OCC policy (all
OCC-capable fileds used in overquaified "where")
By "all OCC-capable fields", you mean all fields that were in the SELECT
clause that originated the SDO data graph?
Today, when DAS generates the update command, only the PK and the fields
being updated will be available on the overqualified WHERE clause. Consider
a table Customer that have the following fields : ID, Lastname, Address; but
we only Lastname is being updated, the overqualified where will only have ID
as it's the table PK, and Lastname being compared to it's original value.
You can see this working on the OCCTest unit test.
The reason I ask this is that I am working in a project developing our own
DAS, and we've had inumerous discussions about what needs to be in the
WHERE clause of an Update stmt in order to avoid conflicts.
For example (I hope the example is not too lame :-) :
Assume a Transaction1 has a Data Graph with SDO objects with employees
data, such as name, salary and a bonus.
T1 changes the SDO by giving a bonus to all employees whose salary are
less than X.
However, a faster Transaction 2 changes the salaries before T1 could
call dataGraph.applyChanges(). Thus making the assumption for the bonus
given by T1 no longer valid.
So, we think that in order to truly avoid conflicts, all selected fields
should be in the WHERE clause of an Update.
Maybe you could make this an option in your DAS config, and based on that,
generate the SQL statement one way or another. Another option is to use a
provided UPDATE command, this would make DAS to use the provided command,
instead of generating one.
Any thoughts?
Thanks a lot, Marilene
> >
> > Hey Ron
>
> > We have talked about it in the past in some threads, and i think we
need
>
> > to bring the subject back now that we have more content available for
a
> DAS
>
> > release. I think we could start by looking into our wiki [1] and
>
> > review/update the list of things we want for DAS M3, and also use [2]
to
>
> > document our release process.
>
> > I'll kick-off the process on the dev-list sometime by end of the week,
>
> > and ask for volunteers for Release Manager.
>
> > [1] -
> http://cwiki.apache.org/confluence/display/TUSCANY/RDB+DAS+-+Releases
>
> > [2] -
http://cwiki.apache.org/confluence/display/TUSCANY/Release+Process
>
> > On 4/10/07, Ron Gavlin <[EMAIL PROTECTED]> wrote:
>
> >
>
> > Java DAS folks,
>
> >
>
> > Are there plans for a Tuscany DAS M3 release to coincide with the
> upcoming
>
> > SDO M3 release? If so, how long after the SDO release do you
anticipate
> the
>
> > DAS release will appear? I suspect the DAS M2 release will be
> incompatible
>
> > with the SDO M3 release, correct?
>
> >
>
> > Thanks in advance for your assistance.
>
> >
>
> > - Ron
>
> >
>
> > ---------------------------------------------------------------------
>
> > To unsubscribe, e-mail: [EMAIL PROTECTED]
>
> > For additional commands, e-mail: [EMAIL PROTECTED]
>
> >
>
> >
>
>
>
> --
>
> Luciano Resende
>
> http://people.apache.org/~lresende <http://people.apache.org/%7Elresende
>
>
--
Luciano Resende
http://people.apache.org/~lresende
