Related to the "Delta Filter" but kept apart for readability

[luissimoes]

I analysed alternatives eg instead of Delta and Sum I used state...
But there is also a problem because the sensor sometimes resets 10 seconds before 00:00 and other times 10 seconds after* 00:00

How can I tell plotly to get always the state for say 30 seconds before now?

I tried

• map_x: x - 30* 1000 but not success

Here the case "before"

[dbuezas]

Not sure what your objective is. Can you clarify?

[luissimoes]

hi
sorry if I was not clear.
this is in part related to the post on the "delta behaviour"...
From the above graph we see that the sensor (its a template as explained in the other post) resets short before 00:00.....

This is a problem when using STATE as statistic and period:day because the last state will be zero... --> wrong results.

To avoid this, since sometimes the sensor resets after 00:00 (as mentioned in the other post, i.e. exactly the symmetrical behaviour relative to 00:00) an alternative would be to use the value say 20 seconds before 00:00

• so I am looking to a way to force a sampling so that it is never at the end of the day but max (say) 20 seconds before

[dbuezas]

Oh, I see. Either map_x before resample or time_offset should both work, but you need to add to x, not subtract

[luissimoes]

Ok 🙏
I still have the discrepancy I mentioned in the previous post (on delta filter) but I will try this.
It seems it only shifts the values…. Doesn’t really resemble them (using offset)

[dbuezas]

What do you mean with "resemble"?

[luissimoes]

Resample 😀

[dbuezas]

Ohhhh i see. Well, add a resample at the end!