|
Re: How to dynamically generate column heading and avoid outof memeory issue [message #794809 is a reply to message #794636] |
Thu, 09 February 2012 18:19 |
|
On issue 1. I do not understand why you would not want to use a direct
query of the db. If this is a restriction that you have to do then that
is fine, but it should not be an issue with BIRT. You can dynamically
generate an entire report using the de api. There are a lot of examples
on birt-exchange.org for this.
On issue 2. Are you using a runandrender task or a run task followed by
a render task. If you use a run task it will write out to an
intermediate file (rptdocument) which can be used by a render task.
Jason
On 2/9/2012 9:28 AM, balachandar sivalingam wrote:
> Hi
>
> For our project we are planning to replace the existing report engine
> with Birts. For that I am doing POC(s). I was told like we should not
> use db query in report instead we should use POJO that will be returned
> from Java Application.
>
> I followed some articles..I used scripted data source with data set. I
> was able to generate the report
>
>
> 1) Now I was told like another restiction is, Field name should be
> dynamically generated that means Columns need to be dynamically
> generated on the runtime.
> Is it possible in birt to do this?
>
> 2) The code i am using is, generateing the report in PDF for all the
> rows at a time and put in the file. What will happen if the row size in
> million.
>
> The Question is:
>
> Can this file be incrementally build in the file system rather than in
> memory. Suppose fetched 300000 rows in db for report. If I extract all
> 300k rows in heap and then build a pdf file again in heap, we would get
> OutOfMemory Exception. Is there any efficient way of doing this in Birt.
> Extracting 1000 rows at a time, open stream , add data and close stream
> and do again.
>
>
> Please clarify on the above.
>
>
> Thanks,
> balachandar
>
|
|
|
Powered by
FUDForum. Page generated in 7.13087 seconds