It has always been best-practice to not trend regular string parameters, for obvious reasons.
I wonder though if the same applies to string parameters of type 'discreet'?
What are the specific features - could be client and/or server side - which are not supported for such parameters?
Should we always aim to convert these to numerical discreets?
What's Skyline's advise on this?
Marieke Goethals [SLC] [DevOps Catalyst] Selected answer as best 19th July 2023
In DataMiner Cube, discreets can all be visualized. They will be evenly distributed on the y axis. Differences with numerical data:
- There is no numeric averaging. This means that the background area for min/max values will be shown across the entire y range from the moment a discreet average point didn't occur across the entire time slot.
- There is no minimum and maximum value for the average points. Instead these contain
- info on how much of the time the average value (the value that occurred the longest) actually occurred (in %)
- What the median value was (aka the value that was set the most times in the time slot)
For string trending, there's an added difference:
- At most, 20 different values will be visualized. If there are any more, they will be categorized on y axis value "(other)".
Toon Casteele [SLC] [DevOps Enabler] Edited answer 15th March 2021