I am using this
SELECT min(“value”) FROM “Haustechnik.Stromzaehler.Momentanleistung_freq” WHERE timeFilter GROUP BY time(__interval) fill(none)
Query. It finds the value 385.
But if I plot the same as a timeseries, I can find manually a value of 325.
The backend is influxdb. What could be the reason for this?
Or I can suggest to use ‘ORDER BY value ASC’ and then choose the first one
GROUP BY time(__interval) fill(none) doesn’t make sense here.
Thank you both.
I used ORDER BY value ASC now and used the lower 1 percentile to exclude outliers (which I had otherwise).
@jangaraj: I agree that it does not make sense, but why does it harm?
The reason why It was there is, as it was there by default.
Default options are good for graph where you need values per time bucket. But you need just single value.
It is waste of resources: there is calculation of local minimums, sorting and returning of x values. But only one value from returned values will be used at the end. It can be very slow in some edge cases.