We found what seems to be a bug when trying to display certain kind of data with the HighDensity Series on the XamDataChart.
Our data has discrete values on Y (3, 5, 7, etc), and many values on X on a small range.
Apparently the High Density Chart can not handle this kind of data, but we would expect a message saying the data is not good for a High Density Chart and not an exception.
Attached you can find a simple Solution reproducing the error.
PS: This is happening with the base version of Infragistics 15.2
Hi,
I have done some investigations why the zoom is not working in the modified project I have provided you in the previous post. It seems that I have commented the progressbar only, but this progressbar is in a grid container that is shown above the xamDataChart. Can you please remove/comment this grid. This will solve the zoom issue.
Thanks,
M.Yovchev
Hi Mircho,
Thanks for your help! The Zoom doesn't seem to work, is that also a consequence of the UseBruteForce flag?
I had guessed after looking at the stack trace that what they found was the cause of the issue. I was hoping to be wrong.
Regarding what the developers said, I don't think there is any circumstance under which you can say "not fixing an Stack Overflow is ok". At the very least they could throw an exception that meant something more clear. Specially considering it works fine to create other Graphs like a Scatter Line or a Bubble Series, although their performance is not good enough, plus they have forced decimation.
For the charts development team, this may be a "degenerate" scenario, but for our users was like the 3rd data-set they tried to load with the High Density Chart, and not something uncommon at all.
Also the data-set of my example is relative small, having only 100k points, we can have data-sets quite bigger. I created that example because I noticed the error happened for that specific data distribution.
In our scenarios it is not feasible to use the "UseBruteForce" flag, because its performance, and we cannot know before hand what kind of data the users will be visualizing, they can switch from one type to other. And it is not possible to infer how its distribution will be.
If they don't want to fix it, at least this post will be useful information for other users with the same issue.
Mircho, thanks again for your help.
Hi Viviana,
I have a response from the development team and they claim that this behavior is not going to be fixed because of a structures that are used and because fixing it will have negative impact on most common cases. But they provided me with a workaround. Here is the answer:
Hello,
I have logged this behavior in our tracking system with ID 212575 and the following case was created for you: CAS-168835-D6G3N5. The next step will be for a developer to review my investigation and confirm my findings or to offer a fix, or other resolution.
Sincerely,
Mircho