We're trying to integrate time series data from an external Web API into our PI System. Our current program design is to search for AF Attributes (PI Point DR) that will receive outputs, then write a rolling time range of data to each attribute (so multiple values per attribute).
I see the PIPoint class has two bulk-write methods - PIPoint.UpdateValues and PIPoint.ReplaceValues, but the AFAttribute class only has AFAttribute.SetValue, which is not a bulk method.
Is there a reason not to write to the attributes' PI Points rather than the attributes themselves? The only reason I can think of is there may be a very small performance penalty for retrieving the attributes' PI Points before the write. I imagine this would be a negligible trade-off for large amounts of attributes and/or data.