tracksolid_timescale_grafana_prod/08_fix_etl_vehicle_key.sql

-- 08_fix_etl_vehicle_key.sql
-- Fixes BUG-01: dwh_gold.refresh_daily_metrics() inserted t.imei (TEXT) into
-- fact_daily_fleet_metrics.vehicle_key (INTEGER REFERENCES dim_vehicles), so
-- every nightly call raised "invalid input syntax for type integer".
--
-- The fix has two parts:
--   1. Seed dwh_gold.dim_vehicles from tracksolid.devices so every IMEI has
--      a serial vehicle_key to point at.
--   2. Rewrite refresh_daily_metrics() to JOIN through dim_vehicles and
--      SELECT the serial key instead of the raw IMEI. The function also
--      upserts dim_vehicles at the top of each run so newly-registered
--      devices appear in the warehouse without manual intervention.

BEGIN;

-- ── 1. Backfill dim_vehicles ─────────────────────────────────────────────────
-- One row per device. is_active mirrors enabled_flag; vehicle_number tracks
-- the plate so dashboards can label charts without joining back to devices.
INSERT INTO dwh_gold.dim_vehicles (imei, vehicle_number, is_active)
SELECT
  d.imei,
  d.vehicle_number,
  COALESCE(d.enabled_flag, 1) = 1
FROM tracksolid.devices d
ON CONFLICT (imei) DO UPDATE SET
  vehicle_number = EXCLUDED.vehicle_number,
  is_active      = EXCLUDED.is_active;

-- ── 2. Rewrite ETL function ──────────────────────────────────────────────────
CREATE OR REPLACE FUNCTION dwh_gold.refresh_daily_metrics(target_date DATE)
RETURNS void LANGUAGE plpgsql AS $$
BEGIN
  -- Sync dim_vehicles first so any IMEI seen in trips has a vehicle_key.
  -- Without this, a brand-new device would have trip rows but no
  -- dim_vehicles entry, and the JOIN below would drop its metrics.
  INSERT INTO dwh_gold.dim_vehicles (imei, vehicle_number, is_active)
  SELECT
    d.imei,
    d.vehicle_number,
    COALESCE(d.enabled_flag, 1) = 1
  FROM tracksolid.devices d
  ON CONFLICT (imei) DO UPDATE SET
    vehicle_number = EXCLUDED.vehicle_number,
    is_active      = EXCLUDED.is_active;

  INSERT INTO dwh_gold.fact_daily_fleet_metrics (
    day,
    vehicle_key,
    total_distance_km,
    total_trips,
    total_drive_hours,
    total_idle_hours,
    fuel_consumed_l,
    alarm_count,
    overspeed_count,
    day_start_time,
    day_end_time,
    avg_speed_kmh,
    peak_speed_kmh
  )
  SELECT
    target_date                                                           AS day,
    dv.vehicle_key                                                        AS vehicle_key,
    ROUND(SUM(t.distance_km)::numeric, 3)                                 AS total_distance_km,
    COUNT(*)                                                              AS total_trips,
    ROUND((SUM(t.driving_time_s) / 3600.0)::numeric, 2)                   AS total_drive_hours,
    ROUND((SUM(t.idle_time_s)    / 3600.0)::numeric, 2)                   AS total_idle_hours,
    ROUND(SUM(t.fuel_consumed_l)::numeric, 3)                             AS fuel_consumed_l,
    COUNT(a.id)                                                           AS alarm_count,
    COUNT(a.id) FILTER (WHERE a.alarm_type ILIKE '%speed%')               AS overspeed_count,
    MIN(t.start_time AT TIME ZONE 'Africa/Nairobi')::TIME                 AS day_start_time,
    MAX(t.end_time   AT TIME ZONE 'Africa/Nairobi')::TIME                 AS day_end_time,
    ROUND(AVG(t.avg_speed_kmh)::numeric, 2)                               AS avg_speed_kmh,
    MAX(t.max_speed_kmh)                                                  AS peak_speed_kmh
  FROM tracksolid.trips t
  JOIN dwh_gold.dim_vehicles dv
    ON dv.imei = t.imei
  LEFT JOIN tracksolid.alarms a
    ON  a.imei = t.imei
    AND DATE(a.alarm_time AT TIME ZONE 'Africa/Nairobi') = target_date
  WHERE DATE(t.start_time AT TIME ZONE 'Africa/Nairobi') = target_date
    AND t.end_time IS NOT NULL
  GROUP BY dv.vehicle_key
  ON CONFLICT (day, vehicle_key) DO UPDATE SET
    total_distance_km = EXCLUDED.total_distance_km,
    total_trips       = EXCLUDED.total_trips,
    total_drive_hours = EXCLUDED.total_drive_hours,
    total_idle_hours  = EXCLUDED.total_idle_hours,
    fuel_consumed_l   = EXCLUDED.fuel_consumed_l,
    alarm_count       = EXCLUDED.alarm_count,
    overspeed_count   = EXCLUDED.overspeed_count,
    day_start_time    = EXCLUDED.day_start_time,
    day_end_time      = EXCLUDED.day_end_time,
    avg_speed_kmh     = EXCLUDED.avg_speed_kmh,
    peak_speed_kmh    = EXCLUDED.peak_speed_kmh;
END;
$$;

COMMENT ON FUNCTION dwh_gold.refresh_daily_metrics(DATE)
  IS 'Populates or refreshes fact_daily_fleet_metrics for the given date. '
     'Joins tracksolid.trips through dwh_gold.dim_vehicles to map IMEI → vehicle_key. '
     'Call nightly: SELECT dwh_gold.refresh_daily_metrics(CURRENT_DATE - 1);';

COMMIT;