Charts: Possible bug in "constant value"

Hi Simen, 

The Warning you receive when using the Constant value function (Function → Signal generator → Constant value), rather than the Constant node, is because it’s being treated like any other function and calculation output – meaning, it’s subject to the 100k data point limit. Therefore, when your time period in view is large enough and/or the granularity you select is low enough, the calculation result will automatically be downsampled.

However, since the value of the constant is always the same value – in most cases – you should be able to ignore that warning. It’s actually best to choose a lower granularity and rely on the automatic data alignment built into the product to resample the data to the relevant granularity based on the other calculations you’ll be performing. So long as you have at least two data points at the start and end of the range in view on the plot, everything in between will be resampled to your constant value of choice. Note: Due to the nature of how our automatic data alignment algorithm handles reindexing, subsequent calculations using this output may not produce values before/after the first/last points (see my comment here for more details).

To that end, it might be better to use the Constant node, as opposed to the Constant value function, since we have some built-in logic to treat these nodes differently (i.e. don’t enforce the downsampling). If you use the Constant node with the Threshold function (Function → Operators → Threshold) to detect if and when a time series crosses a given value, the output will automatically resample at the proper granularity (so long as the time series has <100k data points for the range in view). 

From there, you can use the Integration function (Function → Operators → Integration) at your time granularity of interest (e.g. 1m) and divide that result by a Constant of the total # of minutes in view and multiply by 100. The final result of this calculation should give you the percentage you’re interested in.

The built-in “Thresholds” feature (along the right-hand sidebar) could also be useful here:

It is possible to use the no-code calculations like a calculator (i.e. all constants, no time series) if you want to divide this time range by the total to get the result as a % of total time. Note: The 100k data point limit also applies for this feature.

Hope this is helpful!