Do not allow inf or nan sensor states in statistics (#55943)

homeassistant/components/sensor/recorder.py

@@ -4,6 +4,7 @@ from __future__ import annotations
 import datetime
 import itertools
 import logging
+import math
 from typing import Callable
 
 from homeassistant.components.recorder import history, statistics
@@ -175,6 +176,14 @@ def _get_units(fstates: list[tuple[float, State]]) -> set[str | None]:
     return {item[1].attributes.get(ATTR_UNIT_OF_MEASUREMENT) for item in fstates}
 
 
+def _parse_float(state: str) -> float:
+    """Parse a float string, throw on inf or nan."""
+    fstate = float(state)
+    if math.isnan(fstate) or math.isinf(fstate):
+        raise ValueError
+    return fstate
+
+
 def _normalize_states(
     hass: HomeAssistant,
     entity_history: list[State],
@@ -189,9 +198,10 @@ def _normalize_states(
         fstates = []
         for state in entity_history:
             try:
-                fstates.append((float(state.state), state))
-            except ValueError:
+                fstate = _parse_float(state.state)
+            except (ValueError, TypeError):  # TypeError to guard for NULL state in DB
                 continue
+            fstates.append((fstate, state))
 
         if fstates:
             all_units = _get_units(fstates)
@@ -221,20 +231,20 @@ def _normalize_states(
 
     for state in entity_history:
         try:
-            fstate = float(state.state)
-            unit = state.attributes.get(ATTR_UNIT_OF_MEASUREMENT)
-            # Exclude unsupported units from statistics
-            if unit not in UNIT_CONVERSIONS[device_class]:
-                if WARN_UNSUPPORTED_UNIT not in hass.data:
-                    hass.data[WARN_UNSUPPORTED_UNIT] = set()
-                if entity_id not in hass.data[WARN_UNSUPPORTED_UNIT]:
-                    hass.data[WARN_UNSUPPORTED_UNIT].add(entity_id)
-                    _LOGGER.warning("%s has unknown unit %s", entity_id, unit)
-                continue
-
-            fstates.append((UNIT_CONVERSIONS[device_class][unit](fstate), state))
+            fstate = _parse_float(state.state)
         except ValueError:
             continue
+        unit = state.attributes.get(ATTR_UNIT_OF_MEASUREMENT)
+        # Exclude unsupported units from statistics
+        if unit not in UNIT_CONVERSIONS[device_class]:
+            if WARN_UNSUPPORTED_UNIT not in hass.data:
+                hass.data[WARN_UNSUPPORTED_UNIT] = set()
+            if entity_id not in hass.data[WARN_UNSUPPORTED_UNIT]:
+                hass.data[WARN_UNSUPPORTED_UNIT].add(entity_id)
+                _LOGGER.warning("%s has unknown unit %s", entity_id, unit)
+            continue
+
+        fstates.append((UNIT_CONVERSIONS[device_class][unit](fstate), state))
 
     return DEVICE_CLASS_UNITS[device_class], fstates
 

tests/components/sensor/test_recorder.py

@@ -1,6 +1,7 @@
 """The tests for sensor recorder platform."""
 # pylint: disable=protected-access,invalid-name
 from datetime import timedelta
+import math
 from unittest.mock import patch
 
 import pytest
@@ -349,6 +350,70 @@ def test_compile_hourly_sum_statistics_amount_reset_every_state_change(
     assert "Error while processing event StatisticsTask" not in caplog.text
 
 
+@pytest.mark.parametrize("state_class", ["measurement"])
+@pytest.mark.parametrize(
+    "device_class,unit,native_unit,factor",
+    [
+        ("energy", "kWh", "kWh", 1),
+    ],
+)
+def test_compile_hourly_sum_statistics_nan_inf_state(
+    hass_recorder, caplog, state_class, device_class, unit, native_unit, factor
+):
+    """Test compiling hourly statistics with nan and inf states."""
+    zero = dt_util.utcnow()
+    hass = hass_recorder()
+    recorder = hass.data[DATA_INSTANCE]
+    setup_component(hass, "sensor", {})
+    attributes = {
+        "device_class": device_class,
+        "state_class": state_class,
+        "unit_of_measurement": unit,
+        "last_reset": None,
+    }
+    seq = [10, math.nan, 15, 15, 20, math.inf, 20, 10]
+
+    states = {"sensor.test1": []}
+    one = zero
+    for i in range(len(seq)):
+        one = one + timedelta(minutes=1)
+        _states = record_meter_state(
+            hass, one, "sensor.test1", attributes, seq[i : i + 1]
+        )
+        states["sensor.test1"].extend(_states["sensor.test1"])
+
+    hist = history.get_significant_states(
+        hass,
+        zero - timedelta.resolution,
+        one + timedelta.resolution,
+        significant_changes_only=False,
+    )
+    assert dict(states)["sensor.test1"] == dict(hist)["sensor.test1"]
+
+    recorder.do_adhoc_statistics(period="hourly", start=zero)
+    wait_recording_done(hass)
+    statistic_ids = list_statistic_ids(hass)
+    assert statistic_ids == [
+        {"statistic_id": "sensor.test1", "unit_of_measurement": native_unit}
+    ]
+    stats = statistics_during_period(hass, zero)
+    assert stats == {
+        "sensor.test1": [
+            {
+                "statistic_id": "sensor.test1",
+                "start": process_timestamp_to_utc_isoformat(zero),
+                "max": None,
+                "mean": None,
+                "min": None,
+                "last_reset": process_timestamp_to_utc_isoformat(one),
+                "state": approx(factor * seq[7]),
+                "sum": approx(factor * (seq[2] + seq[3] + seq[4] + seq[6] + seq[7])),
+            },
+        ]
+    }
+    assert "Error while processing event StatisticsTask" not in caplog.text
+
+
 @pytest.mark.parametrize(
     "device_class,unit,native_unit,factor",
     [