Wrong value on aggregation (daily consumption)

[fgottschalck]

Hey all,
I know this is a frequently asked question, but I seem to just do not get it right. I want the daily power consumption.
This is what I build so far:

from(bucket: "iobroker")
  |> range(start: -1d)
  |> filter(fn: (r) => r["_measurement"] == "Wechselrichter.Ertrag.Total.Summe")
  |> filter(fn: (r) => r["_field"] == "value")
  |> aggregateWindow(every: 1d, fn: last, timeSrc:"_start")
  |> difference()

Looking sharp as:

But: the value is not correct. When I go into the data source and look at the values it should be 2,7 kWh, not 2,3 kWh.

Anyone with a hint for me whar could be wrong?

Thanks

[infofcc3]

aggregateWindow() function | Flux Documentation

try timeSrc: "_time", not “_start”

[fgottschalck]

@infofcc3

_time

Thanks - this results in “No Data” - Error: column “_time” doesnt exist

[grant2]

What about timeSrc:“_stop”

[fgottschalck]

@fgottschalck

from(bucket: "iobroker")
  |> range(start: -1d)
  |> filter(fn: (r) => r["_measurement"] == "Wechselrichter.Ertrag.Total.Summe")
  |> filter(fn: (r) => r["_field"] == "value")
  |> aggregateWindow(every: 1d, fn: last, timeSrc:"_start")
  |> difference()

Thanks @grant2 - unfortuntately same - wrong - result.
Interestingly: when I am doing the same with monthly values, it calculates correct values.

[grant2]

Did you read thru this thread?

New to Influxdb Grafana query

@myozone difference() returns the difference between subsequent values. You are querying a whole day’s worth of readings, so you need to use the spread() function as previously described. Also, there is a way in Flux to specify yesterday (Grafana too, but let’s focus on Flux). Something like this should work: import "experimental/date/boundaries" day = boundaries.yesterday() from(bucket: "buckett") |> range(start: day.start, stop: day.stop) |> filter(fn: (r) => r["_measurement"] == "kwh…

[fgottschalck]

Thanks again @grant2 - did read this now, but to be honest, do not understand ist. I fiddled around with ChatGPT too but still no exakt data. I think I will go the way to catch a value via Blockly/javascript at midnight, this seems to be easier for a newbie like me.
Thanks again for your help.