Tag history queries are kind of messed up

Hello folks,

I’m trying to figure out what’s going on with tag history. I’m on 8.1.10.
The results for pointCount and periodic queries are utterly wrong. I’m comparing the results of the different modes and here’s what I find:
asStored: I get about 330 points, with values that make sense

pointCount, set on 500 points:

pointCount starts looking like asStored when configured on > ~12000 points:

periodic looks exactly like pointCount when set for intervals above ~45 seconds, and behaves properly for shorter intervals.

This is quite upsetting when using powerChart, as I’d need to set it to ~12000 points to get the proper display, when it seems about 330 points are enough, as shown by asStored !

What’s up with that ? Am I missing something ? Can someone explain to me what’s going on or how to fix it ?

Thank you in advance, and have a wonderful day !

Edit: So it appears I cannot put images in a <details> block. Huh.

This sounds like a problem with the interpolation the power chart is doing, which from what i understand is one of the frustrations with power charts.

Your tag history is probably fine, what do you get if you use a time series chart and a tag history data source?

The first image I posted. This was as a reference, to show that the problem doesn’t seem to lie with the tag history itself, but with how it’s queried.
So, I can use a time series chart or an XY chart to display data properly, but it’s quite frustrating when there’s the power chart that would be so convenient… I had to resort to making my own component, based on one of the two previously mentioned charts, dropdowns and scripts, to display data. That sucks.

1 Like

No argument here. You are far from the fist to run into this and come out disatisfied, but good news, they are working on it.

Yes that’s good news.
But I’d still like to understand what’s going on with the pointCount and periodic queries… There must be something very weird happening. The second chart shows that quite clearly, the data just gets ‘flattened’ at some point by the aggregation or something, to a constant value, that has absolutely nothing to do with the real data that gets pulled from the historian…