yesterday's sum for is incorrect Topic is solved

[HansR]

Version: v2026.1
Platform: RAspberry Pi 3B+
hardware: P1 Smart meter USB
Description: In the log for the gas the sum of yesterday is what it should be. See chart below: 13 april as example, chart is of 14 april 7h15. 13 april seems to show the value of the 12th.

[gizmocuz]

What happens when you press report and click on todays month

[HansR]

There must be days of zero gas. So the first chart does not chart the report numbers which seem correct.

[waltervl]

It looks like spike detection (introduced in 2026.1) has been activated on this graph?
There is now a function in the graph section that detects 0 values and thinks the meter is incorrect and that the value with data is a total of the last days with 0 and flatten that value out over the last days with 0.

[HansR]

It looks like spike detection (introduced in 2026.1) has been activated on this graph?
There is now a function in the graph section that detects 0 values and thinks the meter is incorrect and that the value with data is a total of the last days with 0 and flatten that value out over the last days with 0.

I am lost.
What do I need to do to get a correct chart?

And btw: I regard this as a bug: you can't introduce something and simply show a wrong chart.

[waltervl]

It looks like spike detection (introduced in 2026.1) has been activated on this graph?
There is now a function in the graph section that detects 0 values and thinks the meter is incorrect and that the value with data is a total of the last days with 0 and flatten that value out over the last days with 0.

I am lost.
What do I need to do to get a correct chart?

And btw: I regard this as a bug: you can't introduce something and simply show a wrong chart.

You cannot get a correct chart at the moment.
And I agree this implementation has some side effects that the implementer did not foresee...
Perhaps someone can create a setting switch that switches this new feature on and off?

[HansR]

You cannot get a correct chart at the moment.

Brilliant

[gizmocuz]

Create a Github issue and mention this forum thread.
A setting seems indeed be the best option. The idea was good, but not when there are valid 0 days, or the algorithm needs to be better.

I assume you are already on the latest beta version and see this behavior? If not, it could already been solved

[imautohuttraeger]

The idea was good, but not when there are valid 0 days, or the algorithm needs to be better.

Word. For example, my small solar power plant can have even several zero days, when there is thick snow on top of the panels. There is even no standby current of the micro-inverter from the grid, as it switches off as long as there is no or not enough power from the solar panels.

[HansR]

I assume you are already on the latest beta version and see this behavior? If not, it could already been solved

Indeed, the current version on the beta channel (v2026.1 (build 17748)) has the problem solved. Thanks.