在 BigQuery 中选择具有不同 2 列的最新事件

Posted 2023-03-24

【中文标题】在 BigQuery 中选择具有不同 2 列的最新事件

【英文标题】：Select the latest event with a distinct 2 columns in BigQuery

【发布时间】：2020-11-11 22:48:00

【问题描述】：

我有一个 BigQuery 表，其架构如下：

  "name": "timeCreated", "type": "datetime",
  "name": "userid", "type": "string",
  "name": "textid", "type": "string",
  "name": "textvalue": "type": "float"

我正在尝试进行查询，因此我最终得到了每对 userid 和 textid 组合的最新 timeCreated 行。我已经尝试过 GROUP BY 等，但我似乎无法通过 timeCreated 字段获取 ORDER，然后删除每对 userid 和 textid 列不在顶部的所有行。

【参考方案1】：

要在 Google BigQuery 中获取组的最新（最后）或最早（第一个）元素，您可以使用 ARRAY_AGG 和 [OFFSET(0)] 和适当的 ORDER BY（DESC 或 ASC）：

WITH test_table AS (
  SELECT DATETIME '2020-11-01 01:00:00' AS timeCreated, 'user1' AS userid, 'text1' AS textid, 1.1 AS textvalue UNION ALL
  SELECT DATETIME '2020-11-01 03:00:00' AS timeCreated, 'user1' AS userid, 'text1' AS textid, 1.2 AS textvalue UNION ALL
  SELECT DATETIME '2020-11-01 02:00:00' AS timeCreated, 'user1' AS userid, 'text1' AS textid, 1.3 AS textvalue UNION ALL
  SELECT DATETIME '2020-11-01 02:00:00' AS timeCreated, 'user1' AS userid, 'text2' AS textid, 1.4 AS textvalue UNION ALL
  SELECT DATETIME '2020-11-01 01:00:00' AS timeCreated, 'user1' AS userid, 'text2' AS textid, 1.5 AS textvalue UNION ALL
  SELECT DATETIME '2020-11-01 00:00:00' AS timeCreated, 'user2' AS userid, 'text1' AS textid, 1.6 AS textvalue
)
SELECT 
  userid,
  textid,
  ARRAY_AGG(timeCreated ORDER BY timeCreated DESC)[OFFSET(0)] AS latest FROM test_table
GROUP BY userid, textid

【参考方案2】：

以下是 BigQuery 标准 SQL

#standardSQL
select as value array_agg(t order by timeCreated desc limit 1)[offset(0)]
from `project.dataset.table` t
group by userid, textid