|
|
|
|
| Re: mean chart question [message #166500 is a reply to message #166460] |
Thu, 01 June 2006 10:27   |
Eclipse User |
|
|
|
Hello David,
Yes, you're right, I meant "columns" by "fields". Let's consider that
SELECT * FROM MY_TABLE would do the trick, actually... I know that a data
model change would rarely occure, but it may in my case.
But I'm probably missing a point, as you mentionned, because my dataset
only has a dynamic number of columns BECAUSE I want to display several
series in my chart.
Is there another way to do it, as we cannot use several datasets, I guess ?
In my case, I'm collecting values set in time for several entities. So I
put the time on X axis and the values on Y, each entity having it's own
columns so to define the series.
But what if the number of entities raises up ?
Am I clear in this example ? My aim would to be to evolve easily with the
report if the model evolves. That may be hopeless.
|
|
|
| Re: mean chart question [message #166569 is a reply to message #166500] |
Thu, 01 June 2006 11:29 |
| Eclipse User |
|
|
|
Thomas,
With BIRT 2.1, you need to define column bindings on each report item,
including the chart. That doesn't leave any place for unknown columns. all
columns have to be defined with this binding, or you won't be able to use
them.
If your DB columns are subject to change it's probably a design issue with
the Database model.
So assuming you're running the chart engine as a standalone library outside
of BIRT, you could write some code to generate Series on the fly in your
chart model based on your columns from the resultset. That would work.
The question now is: are you using BIRT reports or Charts as a standlaone
library?
Thanks,
David
"Thomas Escolan" <thomas_escolan@yahoo.fr> wrote in message
news:f1980fa21de85e52cc8c0ca17d9f225a$1@www.eclipse.org...
> Hello David,
>
> Yes, you're right, I meant "columns" by "fields". Let's consider that
> SELECT * FROM MY_TABLE would do the trick, actually... I know that a data
> model change would rarely occure, but it may in my case.
>
> But I'm probably missing a point, as you mentionned, because my dataset
> only has a dynamic number of columns BECAUSE I want to display several
> series in my chart.
> Is there another way to do it, as we cannot use several datasets, I guess
> ?
>
> In my case, I'm collecting values set in time for several entities. So I
> put the time on X axis and the values on Y, each entity having it's own
> columns so to define the series.
> But what if the number of entities raises up ?
>
> Am I clear in this example ? My aim would to be to evolve easily with the
> report if the model evolves. That may be hopeless.
>
|
|
|
| Re: mean chart question [message #166871 is a reply to message #166569] |
Fri, 02 June 2006 04:26 |
| Eclipse User |
|
|
|
Thanks for your replies, David,
For the moment, I'm not using libraries, but planning to : I'm evaluating
BIRT for a quite heavy chart display web app, and several screens use the
same charts in combination with others (thus I'll need the library support
to code the charts only once).
Your xml report dynamic generating sounds like a VERY GOOD idea, actually
; and I'll recall this later if needed. I have to bypass my "owner file
coding" habits to take benefit of Open Source projects !
To others interested : see the recent thread about Excel translation with
XSL : that's great and hopefull to future issuers.
|
|
|
Powered by
FUDForum. Page generated in 0.04835 seconds
Home