further improve accuracy in low precision mode and low data sample count

grafana/dashboards/statistics.json

@@ -823,7 +823,7 @@
           "format": "table",
           "hide": false,
           "rawQuery": true,
-          "rawSql": "-- This query is used to calculate \"Ø Consumption (gross)\" in Statistics and Trips (with minor changes).\n-- Please make sure to apply changes in both places!\n\nwith gen_date_series as (\n\n  select \n    generate_series(\n      date_trunc('$period', to_timestamp(${__from:date:seconds})),\n      date_trunc('$period', to_timestamp(${__to:date:seconds})),\n      interval '1 $period'\n    ) as series_id\n\n),\n\ndate_series as (\n\n  select\n    greatest(series_id, to_timestamp(${__from:date:seconds})) at time zone 'UTC' as series_id,\n    -- before joining, get beginning of next series to be able to left join `charging_processes`\n    least(lead(series_id) over (order by series_id asc), to_timestamp(${__to:date:seconds})) at time zone 'UTC' as next_series_id\n  from gen_date_series\n  where series_id >= (select date_trunc('$period', p.\"date\" at time zone 'UTC') from positions p where car_id = $car_id and ideal_battery_range_km is not null order by date asc limit 1)\n\n),\n\nbuckets_x_charging_processes as (\n\n  select\n    ds.*,\n    coalesce(cp.car_id, $car_id) as car_id,\n    lag(cp.end_${preferred_range}_range_km) over w as start_range_km,\n    lag(cp.start_date) over w as start_date,\n    lag(p.odometer) over w as start_odometer,\n    case\n      when lag(cp.end_${preferred_range}_range_km) over w is null then null\n      else cp.start_${preferred_range}_range_km\n    end as end_range_km,\n    case\n      when lag(cp.end_date) over w is null then null\n      else cp.start_date\n    end as end_date,\n    case\n      when lag(p.odometer) over w is null then null\n      else p.odometer\n    end as end_odometer,\n    count(start_date) over w as charging_processes_count\n  from date_series ds\n    left join charging_processes cp on ds.series_id <= cp.start_date and ds.next_series_id >= cp.end_date\n    left join positions p on cp.position_id = p.id\n  where cp.car_id = $car_id or cp.car_id is null\n  window w as (partition by ds.series_id order by cp.start_date)\n\n),\n\nbuckets_incl_charging_processes as (\n\n  select\n    series_id,\n    next_series_id,\n    car_id,\n    coalesce(sum(start_range_km - end_range_km), 0) as range_loss,\n    coalesce(max(end_odometer) - min(start_odometer), 0) as distance,\n    coalesce(min(start_date), next_series_id) as start_date,\n    coalesce(max(end_date), next_series_id) as end_date\n  from buckets_x_charging_processes\n  where start_range_km is not null or charging_processes_count < 2\n  group by 1, 2, 3\n\n),\n\nbuckets_incl_charging_processes_x_surrounding_data as (\n\n  select\n    b_incl_cp.*,\n    date_trunc('$period', timezone('UTC', series_id)) as date,\n    lag(coalesce(d.start_${preferred_range}_range_km, p.${preferred_range}_battery_range_km)) OVER w - coalesce(d.start_${preferred_range}_range_km, p.${preferred_range}_battery_range_km) AS range_loss_b_and_a_cp,\n    coalesce(d.start_km, p.odometer) - lag(coalesce(d.start_km, p.odometer)) OVER w AS distance_b_and_a_cp\n  from buckets_incl_charging_processes b_incl_cp\n    left join drives d on\n      0 = $high_precision\n      and d.car_id = b_incl_cp.car_id\n      and (\n        d.start_date >= b_incl_cp.series_id\n        and d.end_date < b_incl_cp.start_date\n        or d.start_date > b_incl_cp.end_date\n        and d.end_date < b_incl_cp.next_series_id\n      )\n    left join positions p on\n      1 = $high_precision\n      and p.ideal_battery_range_km is not null\n      and p.car_id = b_incl_cp.car_id\n      and (\n        p.\"date\" >= b_incl_cp.series_id\n        and p.\"date\" < b_incl_cp.start_date\n        or p.\"date\" > b_incl_cp.end_date\n        and p.\"date\" <= b_incl_cp.next_series_id\n      )\n  window w as (partition by b_incl_cp.series_id, coalesce(d.start_date, p.\"date\") >= b_incl_cp.end_date ORDER BY coalesce(d.start_date, p.\"date\"))\n    \n)\n\nselect\n  EXTRACT(EPOCH FROM date)*1000 AS date_from,\n  EXTRACT(EPOCH FROM date + interval '1 $period')*1000 AS date_to,\n  CASE '$period'\n    WHEN 'month' THEN to_char(date, 'YYYY Month')\n    WHEN 'year' THEN to_char(date, 'YYYY')\n    WHEN 'week' THEN 'week ' || to_char(date, 'WW') || ' starting ' || to_char(date, 'YYYY-MM-DD')\n    ELSE to_char(date, 'YYYY-MM-DD')\n  END AS display,\n  date,\n  ((range_loss + coalesce(sum(range_loss_b_and_a_cp), 0)) * c.efficiency * 1000) / nullif(convert_km((distance + coalesce(sum(distance_b_and_a_cp), 0))::numeric, '$length_unit'), 0) as consumption_gross_$length_unit\nfrom buckets_incl_charging_processes_x_surrounding_data\n  inner join cars c on car_id = c.id\n  where range_loss_b_and_a_cp >= 0 and distance_b_and_a_cp = 0 or distance_b_and_a_cp > 0 or range_loss_b_and_a_cp is null and distance_b_and_a_cp is null\n  group by 1, 2, 3, 4, range_loss, distance, c.efficiency\n  order by date desc\n",
+          "rawSql": "-- This query is used to calculate \"Ø Consumption (gross)\" in Statistics and Trips (with minor changes).\n-- Please make sure to apply changes in both places!\n\nwith gen_date_series as (\n\n  select \n    generate_series(\n      date_trunc('$period', to_timestamp(${__from:date:seconds})),\n      date_trunc('$period', to_timestamp(${__to:date:seconds})),\n      interval '1 $period'\n    ) as series_id\n\n),\n\ndate_series as (\n\n  select\n    greatest(series_id, to_timestamp(${__from:date:seconds})) at time zone 'UTC' as series_id,\n    -- before joining, get beginning of next series to be able to left join `charging_processes`\n    least(lead(series_id) over (order by series_id asc), to_timestamp(${__to:date:seconds})) at time zone 'UTC' as next_series_id\n  from gen_date_series\n  where series_id >= (select date_trunc('$period', p.\"date\" at time zone 'UTC') from positions p where car_id = $car_id and ideal_battery_range_km is not null order by date asc limit 1)\n\n),\n\nbuckets_x_charging_processes as (\n\n  select\n    ds.*,\n    coalesce(cp.car_id, $car_id) as car_id,\n    -- used to skip drives / positions between first charging process start & last charging process end\n    first_value(cp.start_date) over w as min_cp_start_date,\n    last_value(cp.end_date) over w as max_cp_end_date,\n    -- used to correctly calculate range loss of last drive / position before first charging process\n    first_value(cp.start_${preferred_range}_range_km) over w as range_before_first_cp,\n    first_value(p.odometer) over w as odometer_before_first_cp,\n    -- used to correctly calculate range loss of first drive / position after last charging process\n    last_value(cp.end_${preferred_range}_range_km) over w as range_after_last_cp,\n    -- building \"drives\" by taking total range loss and distance between between charging processes\n    lag(cp.end_${preferred_range}_range_km) over w as start_range_km,\n    lag(p.odometer) over w as start_odometer,\n    case\n      when lag(cp.end_${preferred_range}_range_km) over w is null then null\n      else cp.start_${preferred_range}_range_km\n    end as end_range_km,\n    case\n      when lag(p.odometer) over w is null then null\n      else p.odometer\n    end as end_odometer,\n    count(start_date) over w as charging_processes_count\n  from date_series ds\n    left join charging_processes cp on ds.series_id <= cp.start_date and ds.next_series_id >= cp.end_date\n    left join positions p on cp.position_id = p.id\n  where cp.car_id = $car_id or cp.car_id is null\n  window w as (partition by ds.series_id order by cp.start_date rows between unbounded preceding and unbounded following)\n\n),\n\nbuckets_incl_charging_processes as (\n\n  select\n    series_id,\n    next_series_id,\n    car_id,\n    -- if no charging process occurred, move dates to end of period so that all drives / positions are used\n    coalesce(min_cp_start_date, next_series_id) as start_date,\n    coalesce(max_cp_end_date, next_series_id) as end_date,\n    -- used to correctly calculate range loss of last drive / position before first charging process\n    range_before_first_cp,\n    odometer_before_first_cp,\n    -- used to correctly calculate range loss of first drive / position after last charging process\n    range_after_last_cp,\n    -- range loss / distance of \"drives\" between charging processes, only available if min. 2 charging processes occurred\n    coalesce(sum(start_range_km - end_range_km), 0) as range_loss,\n    coalesce(max(end_odometer) - min(start_odometer), 0) as distance\n  from buckets_x_charging_processes\n  where start_range_km is not null or charging_processes_count < 2\n  group by 1, 2, 3, 4, 5, 6, 7, 8\n\n),\n\nbuckets_incl_charging_processes_x_surrounding_data as (\n\n  select\n    b_incl_cp.*,\n    date_trunc('$period', timezone('UTC', series_id)) as date,\n    case\n      -- if first drive / position after last charging process\n      when coalesce(d.start_date, p.\"date\") = first_value(coalesce(d.start_date, p.\"date\")) over w \n        and coalesce(d.start_date, p.\"date\") >= b_incl_cp.end_date\n      -- use range after last charging process\n      then range_after_last_cp\n      -- else use (start) range of drive / position\n      else coalesce(d.start_${preferred_range}_range_km, p.${preferred_range}_battery_range_km)\n    end - coalesce(\n      -- subtract (start) range of next drive / position\n      lead(coalesce(d.start_${preferred_range}_range_km, p.${preferred_range}_battery_range_km)) over w,\n      -- in case of last drive / position before first charging process, use range before charging process\n      case\n        when coalesce(d.start_date, p.\"date\") < b_incl_cp.start_date\n        then range_before_first_cp\n        else null\n      end\n    ) AS range_loss_b_and_a_cp,\n    coalesce(\n      -- (start) odometer of next drive / position\n      lead(coalesce(d.start_km, p.odometer)) over w,\n      -- in case of last drive / position before first charging process, use odometer of charging process\n      case\n        when coalesce(d.start_date, p.\"date\") < b_incl_cp.start_date\n        then odometer_before_first_cp\n        else null\n      end\n    ) - coalesce(d.start_km, p.odometer) AS distance_b_and_a_cp\n  from buckets_incl_charging_processes b_incl_cp\n    left join drives d on\n      0 = $high_precision\n      and d.car_id = b_incl_cp.car_id\n      and (\n        d.start_date >= b_incl_cp.series_id\n        and d.end_date < b_incl_cp.start_date\n        or d.start_date > b_incl_cp.end_date\n        and d.end_date < b_incl_cp.next_series_id\n      )\n    left join positions p on\n      1 = $high_precision\n      and p.ideal_battery_range_km is not null\n      and p.car_id = b_incl_cp.car_id\n      and (\n        p.\"date\" >= b_incl_cp.series_id\n        and p.\"date\" < b_incl_cp.start_date\n        or p.\"date\" > b_incl_cp.end_date\n        and p.\"date\" <= b_incl_cp.next_series_id\n      )\n  window w as (partition by b_incl_cp.series_id, coalesce(d.start_date, p.\"date\") >= b_incl_cp.end_date ORDER BY coalesce(d.start_date, p.\"date\"))\n    \n)\n\nselect\n  EXTRACT(EPOCH FROM date)*1000 AS date_from,\n  EXTRACT(EPOCH FROM date + interval '1 $period')*1000 AS date_to,\n  CASE '$period'\n    WHEN 'month' THEN to_char(date, 'YYYY Month')\n    WHEN 'year' THEN to_char(date, 'YYYY')\n    WHEN 'week' THEN 'week ' || to_char(date, 'WW') || ' starting ' || to_char(date, 'YYYY-MM-DD')\n    ELSE to_char(date, 'YYYY-MM-DD')\n  END AS display,\n  date,\n  ((range_loss + coalesce(sum(range_loss_b_and_a_cp), 0)) * c.efficiency * 1000) / nullif(convert_km((distance + coalesce(sum(distance_b_and_a_cp), 0))::numeric, '$length_unit'), 0) as consumption_gross_$length_unit\nfrom buckets_incl_charging_processes_x_surrounding_data\n  inner join cars c on car_id = c.id\n  where range_loss_b_and_a_cp >= 0 and distance_b_and_a_cp = 0 or distance_b_and_a_cp > 0 or range_loss_b_and_a_cp is null and distance_b_and_a_cp is null\n  group by 1, 2, 3, 4, range_loss, distance, c.efficiency\n  order by date desc\n",
           "refId": "D",
           "sql": {
             "columns": [