Question

我一直在尝试在Bigquery中向上和向下舍入到NEAREST分钟。有谁知道实现此目标的最佳功能和方法？

user_id     |  created_at
-------------------------------------
14451       | 2019-01-31 04:51:28 UTC
14452       | 2019-01-31 04:51:31 UTC
14453       | 2019-01-31 04:51:59 UTC
14454       | 2019-01-31 04:51:03 UTC

我想要的结果如下

user_id     |  created_at
-------------------------------------
14451       | 2019-01-31 04:51:00 UTC
14452       | 2019-01-31 04:52:00 UTC
14453       | 2019-01-31 04:52:00 UTC
14454       | 2019-01-31 04:51:00 UTC

非常感谢您的帮助。

Answer 1

以下是用于BigQuery标准SQL

#standardSQL
SELECT user_id, 
  TIMESTAMP_TRUNC(TIMESTAMP_ADD(created_at, INTERVAL 30 SECOND), MINUTE) created_at
FROM `project.dataset.table`

您可以使用问题中的示例数据来测试，玩这个游戏，如以下示例所示

#standardSQL
WITH `project.dataset.table` AS (
  SELECT 14451 user_id, TIMESTAMP '2019-01-31 04:51:28 UTC' created_at UNION ALL
  SELECT 14452, '2019-01-31 04:51:31 UTC' UNION ALL
  SELECT 14453, '2019-01-31 04:51:59 UTC' UNION ALL
  SELECT 14454, '2019-01-31 04:51:03 UTC' 
)
SELECT user_id, 
  TIMESTAMP_TRUNC(TIMESTAMP_ADD(created_at, INTERVAL 30 SECOND), MINUTE) created_at
FROM `project.dataset.table`

有结果

Row user_id created_at   
1   14451   2019-01-31 04:51:00 UTC  
2   14452   2019-01-31 04:52:00 UTC  
3   14453   2019-01-31 04:52:00 UTC  
4   14454   2019-01-31 04:51:00 UTC

Answer 2

在执行DIV之前，只需将60的一半（值30）相加

SELECT user_id ,SEC_TO_TIME(((TIME_TO_SEC(created_at)+30) DIV 60) * 60) FROM `table`

Bigquery：有没有办法将时间戳上移或下移到NEAREST分钟？

2 个答案: