Learning Forum

[AndyJP]

I am plotting a parameter relation between two parameters found in optimization. Intermadiate data is erased, and the parameters are clean-simulated again with frequency sweep. In the end I have 7 converged "last passes", and 7 corresponding sweeps.

When I plot a modal rectangular plot for "last adaptive" with X and Y representing both parameters, I can get a single point for every result from the set of 7, separately, by manually setting one of the parameters.

But when I select plotting "ALL" parameters at once, two data points are lost! There is no visible rule, which data points are lost, but at the moment I miss point #3 and point #7 of 7.

However, when I plot "sweep", I see all the points.

Yes, I know that I can keep plotting sweep for parameter combination plots. But that is clearly a bug, which made me wonder in guess, wheare are the absent points for some meonths. And may make harm to other projects, where people do not expect such behaviour of the report engine.

[AndyJP]

I think, I got the Idea where they gone. I was using "Hide" for the legend and some "adaptive pass#" points a few weeks ago. That data has gone and overwritten many many times. The plot itself was changed with different X&Y
But for some reason, when the plot is changed, the hidden attribute of the deleted data reappears on new unrelated computation points in different (last adaptive) mode. Quite inconvenient and dangerous behavior.

[AndyJP]

Ah, an yes, some phanton non-existent data points appear on the "last adaptive" plot, probably belonging to non-converged passes, when I unhide all.