-- 埋点 raw 层 INSERT 冒烟（ES SparkSQL 直读 → Hive ORC）
-- 跑：python3 bin/spark-sql-starter.py -f tests/integration/tracking/es_insert_smoke.sql
-- 前置：
--   1. 集群 spark/jars/ 已挂 elasticsearch-spark-20_2.11-7.17.29.jar（kb/01 §5）
--   2. 已建表：python3 bin/spark-sql-starter.py -f tests/integration/tracking/ddl/hive_raw_es.sql

CREATE TEMPORARY VIEW v_es_traces
USING org.elasticsearch.spark.sql
OPTIONS (
  'es.nodes' = '192.168.33.21',
  'es.port' = '9200',
  'es.nodes.wan.only' = 'true',
  'es.resource' = 'traces-2026-04-08'
);

INSERT OVERWRITE TABLE test.raw_usr_traces_apd_d_es PARTITION (dt='20260408')
SELECT
    _class,
    anonymousId,
    distinctId,
    event,
    CAST(flushTime AS STRING) AS flushTime,
    to_json(lib) AS lib,
    loginId,
    CAST(productInstock AS STRING) AS productInstock,
    to_json(properties) AS properties,
    CAST(`time` AS STRING) AS `time`,
    type
FROM v_es_traces LIMIT 5000;

SELECT COUNT(*) AS cnt FROM test.raw_usr_traces_apd_d_es WHERE dt='20260408';

SELECT _class, anonymousId, event, `time`,
       get_json_object(properties, '$.os') AS os,
       get_json_object(properties, '$.params.groupId') AS group_id
FROM test.raw_usr_traces_apd_d_es WHERE dt='20260408' LIMIT 3;