chore: remove duplicates from N plates before and after

queries/models/utils/pedido_n_placas_antes_depois.sql

@@ -1,7 +1,7 @@
 DECLARE plate STRING DEFAULT '';
-DECLARE start_datetime DATETIME DEFAULT '2024-06-15T17:12:18';
-DECLARE end_datetime DATETIME DEFAULT '2024-06-15T17:15:52';
-DECLARE N INT64 DEFAULT 1;
+DECLARE start_datetime DATETIME DEFAULT '2024-01-01T00:00:00';
+DECLARE end_datetime DATETIME DEFAULT  '2024-01-01T00:00:00';
+DECLARE N INT64 DEFAULT 20;
 
 WITH all_readings AS (
  SELECT
@@ -14,7 +14,7 @@ WITH all_readings AS (
  camera_latitude AS latitude,
  camera_longitude AS longitude,
  DATETIME(datahora_captura, 'America/Sao_Paulo') AS datahora_captura,
- ROW_NUMBER() OVER (PARTITION BY camera_numero ORDER BY datahora) AS row_num
+ ROW_NUMBER() OVER (PARTITION BY placa, datahora ORDER BY datahora) AS row_num_duplicate
  FROM `rj-cetrio.ocr_radar.readings_*`
  WHERE
  DATETIME(datahora, "America/Sao_Paulo")
@@ -23,6 +23,22 @@ WITH all_readings AS (
  ORDER BY datahora
 ),
 
+clean_reading AS (
+ SELECT
+ placa,
+ tipoveiculo,
+ velocidade,
+ datahora_local,
+ camera_numero,
+ empresa,
+ latitude,
+ longitude,
+ datahora_captura,
+ ROW_NUMBER() OVER (PARTITION BY camera_numero ORDER BY datahora_local) AS row_num
+ FROM all_readings
+ WHERE row_num_duplicate = 1
+),
+
 selected AS (
  SELECT
  placa,
@@ -35,7 +51,7 @@ selected AS (
  longitude,
  datahora_captura,
  row_num AS selected_row_num,
- FROM all_readings
+ FROM clean_reading
  WHERE placa = plate
  AND datahora_local
  BETWEEN start_datetime
@@ -45,7 +61,7 @@ selected AS (
 before_and_after AS (
  SELECT
  a.*
- FROM all_readings a
+ FROM clean_reading a
  JOIN selected s
  ON a.camera_numero = s.camera_numero
  AND (a.row_num BETWEEN s.selected_row_num - N AND s.selected_row_num + N)