Correct statistic history for _trip_total_ entities

[virtualdj]

Brief description

I think every entity that has _last_trip_total_ in its name should also set the last_reset attribute to let Home Assistant calculate a proper "total" statistic.

Breakdown

While I was exporting my last trip sensor data from Home Assistant, I notied something weird.
This is an example:

statistic_id	unit	start	sum	state	delta
sensor.golf_last_trip_total_electric_consumption	kWh	07.01.2026 12:00	0.171	1.476	0
sensor.golf_last_trip_total_electric_consumption	kWh	07.01.2026 13:00	0.171	1.476	0
sensor.golf_last_trip_total_electric_consumption	kWh	07.01.2026 14:00	0.171	1.476	0
sensor.golf_last_trip_total_electric_consumption	kWh	07.01.2026 15:00	0.171	1.476	0
sensor.golf_last_trip_total_electric_consumption	kWh	07.01.2026 16:00	-0.495	0.81	-0.666
sensor.golf_last_trip_total_electric_consumption	kWh	07.01.2026 17:00	-0.495	0.81	0
sensor.golf_last_trip_total_electric_consumption	kWh	07.01.2026 18:00	-0.495	0.81	0
sensor.golf_last_trip_total_electric_consumption	kWh	07.01.2026 19:00	-0.495	0.81	0
sensor.golf_last_trip_total_electric_consumption	kWh	07.01.2026 20:00	-0.495	0.81	0
sensor.golf_last_trip_total_electric_consumption	kWh	07.01.2026 21:00	-0.495	0.81	0
sensor.golf_last_trip_total_electric_consumption	kWh	07.01.2026 22:00	-0.495	0.81	0
sensor.golf_last_trip_total_electric_consumption	kWh	07.01.2026 23:00	-0.495	0.81	0
sensor.golf_last_trip_total_electric_consumption	kWh	08.01.2026 00:00	-0.495	0.81	0
sensor.golf_last_trip_total_electric_consumption	kWh	08.01.2026 01:00	-0.495	0.81	0
sensor.golf_last_trip_total_electric_consumption	kWh	08.01.2026 02:00	-0.495	0.81	0
sensor.golf_last_trip_total_electric_consumption	kWh	08.01.2026 03:00	-0.495	0.81	0
sensor.golf_last_trip_total_electric_consumption	kWh	08.01.2026 04:00	-0.495	0.81	0
sensor.golf_last_trip_total_electric_consumption	kWh	08.01.2026 05:00	-0.495	0.81	0
sensor.golf_last_trip_total_electric_consumption	kWh	08.01.2026 06:00	0.225	1.53	0.72
sensor.golf_last_trip_total_electric_consumption	kWh	08.01.2026 07:00	0.225	1.53	0
sensor.golf_last_trip_total_electric_consumption	kWh	08.01.2026 08:00	0.225	1.53	0
sensor.golf_last_trip_total_electric_consumption	kWh	08.01.2026 09:00	0.225	1.53	0
sensor.golf_last_trip_total_electric_consumption	kWh	08.01.2026 10:00	0.225	1.53	0
sensor.golf_last_trip_total_electric_consumption	kWh	08.01.2026 11:00	0.225	1.53	0

While the chart is correct (notice the second to last column, state, that corresponds to the chart values):

I don't think the data above is, too! The third to last column is the sum and I don't think it is correct to have a negative number (-0.495) as I've not produced energy. The delta is wrong indeed.

The problem is that the VW backend provides the last trip data only when it has finished, i.e. it is not a real energy counter. According to the Home Assistant documentation here and here, I think that last_trip_total_xxx_consumption sensors should also set the last_reset attribute when their state value has changed. This should hopefully provide a proper statistics data, while the chart values should not change. What do you think?

Long-term & refuel totals

This certainly does not apply to the longterm_trip_total_ and refuel_trip_total_ sensors, because they reset within a larger period. However, from the examples of the HA documentation, probably they should be set to total_increasing instead to sum the delta when they reset to 0?

[virtualdj]

I'm still reflecting on this topic - when spare time allows - and I believe the correct way to handle those is value is the following.

Last trip

Affected sensors

last_trip_total_electric_consumption
last_trip_total_fuel_consumption

Logic

They should use "state_class": VWStateClass.TOTAL as they are now, but the integration should also set last_reset to the date they changed. This is because:

• measurement doesn't support counting the energy (kWh) or volume (L)

• total_increasing is not OK, because if we have 3 trips like below it doesn't always sum them:

  Trip #	state	sum	note
  1	10 kWh	10	OK, first 10 added
  2	12 kWh	2	WRONG, it is a new trip, so 12 should be added!
  3	5 kWh	7	OK, 5 has been added as it detected the zero-cross

• total alone is still not OK, because in addition to the missing sum of the previous example it also adds negative values:

  Trip #	state	sum	note
  1	10 kWh	10	OK, first 10 added
  2	12 kWh	2	WRONG, it is a new trip, so 12 should be added!
  3	5 kWh	-7	WRONG, it should not decrease as it's energy/volume consumed!

• total with setting last_reset each time to a different value when state changes should be the solution to have something like this:

  Trip #	state	last_reset	sum
  1	10 kWh	date_of_trip_1	10
  2	12 kWh	date_of_trip_2	22
  3	5 kWh	date_of_trip_3	27

Possible code fix

In the Home Assistant integration __init__.py change to:

@callback
def async_write_ha_state(self) -> None:
    """Write state to HA, but only if needed."""
    try:
        backend_refresh_time = self.instrument.last_refresh
    except ValueError:
        # Instrument not found - mark as unavailable
        _LOGGER.debug(
            "Instrument not found for entity %s, marking as unavailable",
            self.entity_id,
        )
        super().async_write_ha_state()
        return

    ############### ADDED PART BEGIN ###############
    # Set the last_reset time for sensors with SensorStateClass.TOTAL
    if (
        hasattr(self, "state_class")
        and self._attr_state_class is not None
        and self._attr_state_class == SensorStateClass.TOTAL
    ):
        self._attr_last_reset = self.instrument.last_refresh
    ################ ADDED PART END ################

    # Get the previous state from the state machine if found
    prev: State | None = self.hass.states.get(self.entity_id)

and add the missing import above:

from homeassistant.components.sensor import SensorStateClass

Refuel and long term trip

Affected sensors:

refuel_trip_total_electric_consumption
refuel_trip_total_fuel_consumption
longterm_trip_total_electric_consumption
longterm_trip_total_fuel_consumption

Logic

These sensors have the same issues of the last trip, because even though they reset after more time they still reset! So the energy/volume sum is calculated incorrectly when this event happens. It is also more difficult to detect the correct last_reset date for these, as the API is not supplying one and we can't just rely on a change of the previous value.

In my opinion they should revert to "state_class": VWStateClass.MEASUREMENT as we're already counting the total energy using the last trip sensor and those would be redundant; but doing so causes a new warning in the log:

Entity sensor.xxxx_refuel_trip_total_electric_consumption (<class 'custom_components.volkswagencarnet.sensor.VolkswagenSensor'>) is using state class 'measurement' which is impossible considering device class ('energy') it is using; expected None or one of 'total', 'total_increasing'; Please update your configuration if your entity is manually configured, otherwise create a bug report at https://github.com/robinostlund/homeassistant-volkswagencarnet/issues
Entity sensor.xxxx_refuel_trip_total_fuel_consumption (<class 'custom_components.volkswagencarnet.sensor.VolkswagenSensor'>) is using state class 'measurement' which is impossible considering device class ('volume') it is using; expected None or one of 'total', 'total_increasing'; Please update your configuration if your entity is manually configured, otherwise create a bug report at https://github.com/robinostlund/homeassistant-volkswagencarnet/issues
Entity sensor.xxxx_long_term_trip_total_electric_consumption (<class 'custom_components.volkswagencarnet.sensor.VolkswagenSensor'>) is using state class 'measurement' which is impossible considering device class ('energy') it is using; expected None or one of 'total', 'total_increasing'; Please update your configuration if your entity is manually configured, otherwise create a bug report at https://github.com/robinostlund/homeassistant-volkswagencarnet/issues
Entity sensor.xxx_long_term_trip_total_fuel_consumption (<class 'custom_components.volkswagencarnet.sensor.VolkswagenSensor'>) is using state class 'measurement' which is impossible considering device class ('volume') it is using; expected None or one of 'total', 'total_increasing'; Please update your configuration if your entity is manually configured, otherwise create a bug report at https://github.com/robinostlund/homeassistant-volkswagencarnet/issues

so "device_class": VWDeviceClass.ENERGY and "device_class": VWDeviceClass.VOLUME should also be removed from them.

Possible code fix

Delete the lines 2190, 2191, 2202, 2203 from here:

volkswagencarnet/volkswagencarnet/vw_dashboard.py

Lines 2182 to 2205 in f2269f3

	(
	Sensor,
	[],
	{
	"attr": "refuel_trip_total_electric_consumption",
	"name": "Refuel trip total electric consumption",
	"icon": "mdi:car-battery",
	"unit": "kWh",
	"device_class": VWDeviceClass.ENERGY,
	"state_class": VWStateClass.TOTAL,
	},
	),
	(
	Sensor,
	[],
	{
	"attr": "refuel_trip_total_fuel_consumption",
	"name": "Refuel trip total fuel consumption",
	"icon": "mdi:fuel",
	"unit": "L",
	"device_class": VWDeviceClass.VOLUME,
	"state_class": VWStateClass.TOTAL,
	},
	),

Delete the lines 2320, 2321, 2332, 2333 from here:

volkswagencarnet/volkswagencarnet/vw_dashboard.py

Lines 2312 to 2335 in f2269f3

	(
	Sensor,
	[],
	{
	"attr": "longterm_trip_total_electric_consumption",
	"name": "Long-term trip total electric consumption",
	"icon": "mdi:car-battery",
	"unit": "kWh",
	"device_class": VWDeviceClass.ENERGY,
	"state_class": VWStateClass.TOTAL,
	},
	),
	(
	Sensor,
	[],
	{
	"attr": "longterm_trip_total_fuel_consumption",
	"name": "Long-term trip total fuel consumption",
	"icon": "mdi:fuel",
	"unit": "L",
	"device_class": VWDeviceClass.VOLUME,
	"state_class": VWStateClass.TOTAL,
	},
	),

I'll try this modification on my system and see if it solves the statistics issue.