[birt-dev] Re: [birt-pmc] parallel data retrieval

[Eric Deshayes <edeshayes@xxxxxxxxx>]

Hi Wenfeng,
thanks for your answer.
I was not aware of such features and I am actually interested in understanding 
how to achieve that.
ow do you set your query to have the second dataset dependant on the first 
one? Is that at the report engine or the data engine level?
I am personally using the BIRT data engine in order to generate live report 
with some realtime update. I am not using the ReportEngine itself but I have 
spent a lot of time diggin into the BIRT code.
Would there be a place where I could find some design and technical 
documentation (more than the javadoc?
Thanks,
Eric
On Friday 16 November 2007 02:10:45 pm Wenfeng Li wrote:
> Eric
>
> Birt-dev@xxxxxxxxxxx is for this type of design and implementation
> discussion.
>
> One reason we are not able to do parallel data fetching is that BIRT
> supports scripting.  Developer can write script such that the visibility
> of the report item that uses  the 2nd data set depends on the value in
> the first data set.  Developer can also write script that changes the
> query in the 2nd dataset based on the value in the first dataset.   All
> this will require BIRT engine to do code analysis before deciding on an
> execution plan.   It would be an interesting project to optimize the
> execution plan of a report based on static analysis of the report design
> XML.
>
>
> Wenfeng
>
>
>
>
>
> -----Original Message-----
> From: birt-pmc-bounces@xxxxxxxxxxx [mailto:birt-pmc-bounces@xxxxxxxxxxx]
> On Behalf Of Eric Deshayes
> Sent: Friday, November 16, 2007 12:05 PM
> To: birt-pmc@xxxxxxxxxxx
> Subject: [birt-pmc] parallel data retrieval
>
>
> Hi all,
> I am not sure this is the best mailing list for such a question.. Let me
> know if there is a better mailing list. At the moment, the data engine
> makes all the data retrieval sequentially. So, if we are accessing two
> independant datasources that respond independantly in the same amount of
> time, we need to wait for both to return some data one after each other.
> Wouldn't that be a good improvement to do the request in parallel, in
> two different thread so the time of the data retrieval will be the time
> of the longest datasource retrieval? I have no idea how feasible it is
> in the current BIRT data engine design. Thanks for your
> feedbacks/comments. Eric _______________________________________________
> birt-pmc mailing list
> birt-pmc@xxxxxxxxxxx https://dev.eclipse.org/mailman/listinfo/birt-pmc
> _______________________________________________
> birt-pmc mailing list
> birt-pmc@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/birt-pmc
-- 
--
Eric Deshayes
edeshayesatgmail.com