ods_usr_traces_apd_d.sql 4.5 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950
  1. -- 作者:tianyu.chu
  2. -- 日期:2026-06-10
  3. -- 工单:(无)
  4. -- 目的:埋点 raw → ods,解析脱敏后 _source JSON 拍平公共属性 + 保留 params_json;按文件日 dt 动态分区写入
  5. -- 状态:[待执行]
  6. -- 备注:日期范围 [${start_date}, ${stop_date}] 含两端、动态分区按 raw.dt 写——日常增量传 start=stop=单日,
  7. -- 历史回算传全区间,一份 SQL 两用(避免解析逻辑复制漂移,对齐 ADR-11);
  8. -- dt = 文件日(N=1 不归位):实测 ~99.4% 事件日=文件日,~0.6% 迟到/未来小偏差按当天落,业务允许(见 kb/13 §8);
  9. -- 动态覆盖本环境只覆盖 SELECT 出现的分区(kb/21 §8);事件不可变,无双源 union / 无去重;时区随集群(东八区)
  10. INSERT OVERWRITE TABLE ods.ods_usr_traces_apd_d PARTITION (dt)
  11. SELECT
  12. es_id AS es_id,
  13. event_name AS event_name,
  14. get_json_object(raw_json, '$.type') AS evt_type,
  15. CAST(CAST(get_json_object(raw_json, '$.time') AS BIGINT) / 1000 AS TIMESTAMP) AS event_time,
  16. CAST(CAST(get_json_object(raw_json, '$.flushTime') AS BIGINT) / 1000 AS TIMESTAMP) AS flush_time,
  17. get_json_object(raw_json, '$.loginId') AS login_id,
  18. get_json_object(raw_json, '$.distinctId') AS distinct_id,
  19. get_json_object(raw_json, '$.anonymousId') AS anonymous_id,
  20. get_json_object(raw_json, '$.properties.userId') AS user_id,
  21. get_json_object(raw_json, '$.properties.userName') AS user_name,
  22. get_json_object(raw_json, '$.properties.userLevel') AS user_lvl,
  23. get_json_object(raw_json, '$.lib.lib') AS lib,
  24. get_json_object(raw_json, '$.lib.libVersion') AS lib_version,
  25. get_json_object(raw_json, '$.lib.libMethod') AS lib_method,
  26. get_json_object(raw_json, '$.properties.appId') AS app_id,
  27. get_json_object(raw_json, '$.properties.appName') AS app_name,
  28. get_json_object(raw_json, '$.properties.appVersion') AS app_version,
  29. get_json_object(raw_json, '$.properties.wgtVersion') AS wgt_version,
  30. get_json_object(raw_json, '$.properties.os') AS os,
  31. get_json_object(raw_json, '$.properties.osVersion') AS os_version,
  32. get_json_object(raw_json, '$.properties.manufacturer') AS manufacturer,
  33. get_json_object(raw_json, '$.properties.brand') AS brand,
  34. get_json_object(raw_json, '$.properties.model') AS model,
  35. get_json_object(raw_json, '$.properties.deviceId') AS device_id,
  36. get_json_object(raw_json, '$.properties.networkType') AS network_type,
  37. get_json_object(raw_json, '$.properties.carrier') AS carrier,
  38. CAST(get_json_object(raw_json, '$.properties.wifi') AS BOOLEAN) AS wifi,
  39. CAST(get_json_object(raw_json, '$.properties.screenWidth') AS BIGINT) AS screen_width,
  40. CAST(get_json_object(raw_json, '$.properties.screenHeight') AS BIGINT) AS screen_height,
  41. CAST(get_json_object(raw_json, '$.properties.timezoneOffset') AS BIGINT) AS timezone_offset,
  42. CAST(get_json_object(raw_json, '$.properties.isFirstDay') AS BOOLEAN) AS is_first_day,
  43. CAST(get_json_object(raw_json, '$.properties.isFirstTime') AS BOOLEAN) AS is_first_time,
  44. CAST(get_json_object(raw_json, '$.properties.resumeFromBackground') AS BOOLEAN) AS resume_from_background,
  45. CAST(get_json_object(raw_json, '$.properties.eventDuration') AS BIGINT) AS event_duration,
  46. get_json_object(raw_json, '$.properties.params') AS params_json,
  47. dt AS dt
  48. FROM raw.raw_usr_traces_apd_d
  49. WHERE dt BETWEEN '${start_date}' AND '${stop_date}';