No longer able to delete invalid temperature log Topic is solved

[WilcoEHV]

Hi,
I've updated from 2024.7 to 2025.1 with Raspberry 3B/ Raspberry OS.

Occasionally get invalid temperature readings with RFXCOM sensors, for example room temperature is all day around 21 degrees but i get 1 reading of 0 degrees.
I delete these daily by clicking on them while holding shift.

With 2025.1 I can delete 1 false reading and than have to restart domoticz to remove another one. This is indepenend of the browser.
Can this be fixed?

And there is also a feature request; please make it possible to delete a false value of an earlier day (shortest log) and have the week, month and year log correct automaticlly.

Kind regards, Wilco

[waltervl]

That is strange. I can remove multiple Temperature readings in one session.

Did you try to start a browser session in private/incognito mode to check if it is a browser cache issue?

Regarding the avarages, 1 incorrect misreading gives an average difference of 0.34% So for an average of 20 it is 0.7 degrees difference. Is your temperature meter so accurate?

You better invest in a better protocol like zigbee or zwave. I never have misreadings with my zigbee temperature devices.

[WilcoEHV]

Hi,
Thx for the reply.
It's browser independant (multiple browsers, multiple systems).

You are right about limited inmpact of a false reading but sometimes there is a interruption for a couple of hours an than the impact is bigger.

The log shows a temp range and that range gets messed up.

[waltervl]

But did you do what I asked for (running the delete command in incognito/private browser session)?

Because if that does not help something is wrong with your environment as it should be possible to delete multiple values. No reboot necessary.

And better get your RF sensors working properly..... Perhaps some other sensor is blocking the temperature one?

[arionp]

I am having the same issue and after some tries I found out I can only delete a single point after each resfresh of the page (hit F5).

Incognito/Privacy mode makes no difference. Also response from server on the delete request returns success:

Code: Select all

{
	"status" : "OK",
	"title" : "deletedatapoint"
}

I can't figure out the difference of the request after the refesh and any subsequent one. They are idnetical (except the data point date of course). Yet the first one works and the rest do not.

A little background for my use case:
I too have some occasional invalid readings (actually zero readings). The problem in my case is that I am using a temperature sensor to monitor an aquarium. I need to make sure the temperature is constant. Since the temp is always around 25C any small deviation (1C-2C) is easily seen in graphs. However when I get a zero reading the graph auto ranges and the rest becomes a flat line, I cannot see small deviations anymore. It's not about accuracy.

[waltervl]

Perhaps check with F12 (Developer tools) what goes wrong. But first be sure you have deleted your browser cache.
I have no issues in deleting multiple temperature values from a day graph.

It could also be you are on an old system as old Safari browsers on old Macbooks also give issues with the graph system.

[waltervl]

I am having the same issue and after some tries I found out I can only delete a single point after each resfresh of the page (hit F5).

Incognito/Privacy mode makes no difference. Also response from server on the delete request returns success:

Code: Select all

{
	"status" : "OK",
	"title" : "deletedatapoint"
}

I can't figure out the difference of the request after the refesh and any subsequent one. They are idnetical (except the data point date of course). Yet the first one works and the rest do not.

After an extra test I see the same behavior. The next delete action is not really executed although it says it does. only after an refresh (F5) of the log page the next delete action will be executed.

[waltervl]

Created Github issue https://github.com/domoticz/domoticz/issues/6341

[waltervl]

Created Github issue https://github.com/domoticz/domoticz/issues/6341

Gizmocuz fixed it in beta 16712 today. So please update and try.

[WilcoEHV]

Hi,
Nice the problem is recognised and fixed!
Hopefully there will be a new stable release soon.
Wilco

[gizmocuz]

No there will not be a new stable soon as it forces users to upgrade their system to OpenSSL3 (Debian bookworm)
Aldo this is a easy step, it's best to keep this for the beta users for now as they are more experienced than users updating to a stable version once in a while.

If your system supports OpenSSL3, or you are running Domoticz in a Docker (compose) environment, then by all means, just upgrade to the beta version as there have not been that much changes since the recept stable

[lamama1234]

No there will not be a new stable soon as it forces users to upgrade their system to OpenSSL3 (Debian bookworm)
Aldo this is a easy step, it's best to keep this for the beta users for now as they are more experienced than users updating to a stable version once in a while.

If your system supports OpenSSL3, or you are running Domoticz in a Docker (compose) environment, then by all means, just upgrade to the beta version as there have not been that much changes since the recept stable

Hi, I just tested this on build 17652 and it still shows the same issue (can delete only one point - the next one will need an F5 refresh).

Thanks

Edit: I added a before and after a "single-point-delete" screenshots. It shows that the x-axis timing changed after deleting that single point and this may explain why the next deletions do not work. After an F5 refresh of the page the x-axis timing sets back to its "good" form and further points deletions are working.

[ressof]

Also have this problem with 17806 release
I delete a point on my rain meter and it disappears but I can not delete a second one. When I press F5 (update) the point I deleted first is back in the log.

[arionp]

The issue is back with version 2026.2. Not sure if it is the same cause.
Looking at the network traces, it seems the first api call after refresh uses the correct datetime, but subsequent calls use datetime offset by 2 hours (presumambly my timezone offset). Also confirmed that on subsequent calls the wrong datapoint is actually deleted.

Edit:
Actually, after the first delete, the times displayed on the chart (both the axis and the tooltip) change to UTC.

Edit2:
Tried both Chrome and Firefox, makes no difference.
Did another test to confirm the hypothesis. Moved mouse over datapoint to be deleted and noted the time, moved the mouse over the datapoint at noted time plus timezone offset, Shift clicked and accepted -> deleted the original datapoint.

[gizmocuz]

Same issue as

viewtopic.php?p=334570