Question

我试图将字符串转换为SQL中的时间戳。问题非常简单，如何将此字符串转换为当天午夜开始的时间戳？

在我的数据库中，我还有一个存储在timestamp_micros中的字段，其中一个可以工作，我认为将微控制器转换为时间戳会比字符串更容易。

例如 20170118 => timestamp

查询：

WITH allTables as (
  SELECT
      event.date as date,
      count(*) as totalSessions,
      count(DISTINCT user_dim.app_info.app_instance_id) as uniqueUsers
  FROM `namehiddenonlyhere.*`
  CROSS JOIN
      UNNEST(event_dim) AS event
  WHERE
      event.name = 'session_start'
  AND 
      event.date <= FORMAT_DATE('%Y%m%d', DATE_SUB(CURRENT_DATE, INTERVAL 1 DAY))
  AND
      event.date >= FORMAT_DATE('%Y%m%d', DATE_SUB(CURRENT_DATE, INTERVAL 7 DAY))
  GROUP BY event.date

  UNION ALL 
  SELECT
      event.date as date,
      count(*) as totalSessions,
      count(DISTINCT user_dim.app_info.app_instance_id) as uniqueUsers
  FROM `namehiddenonlyhere.*`
  CROSS JOIN
      UNNEST(event_dim) AS event
  WHERE
      event.name = 'session_start'
  AND 
      event.date <= FORMAT_DATE('%Y%m%d', DATE_SUB(CURRENT_DATE, INTERVAL 1 DAY))
  AND
      event.date >= FORMAT_DATE('%Y%m%d', DATE_SUB(CURRENT_DATE, INTERVAL 7 DAY))
  GROUP BY event.date
) SELECT 'all apps' as target, date as datapoint_time, totalSessions datapoint_value FROM allTables;

Answer 1

您可以使用PARSE_DATE或PARSE_TIMESTAMP（格式字符串相同）来获取日期或时间戳。例如：

SELECT
  d,
  PARSE_DATE('%Y%m%d', d) AS date,
  PARSE_TIMESTAMP('%Y%m%d', d) AS timestamp
FROM UNNEST(['20170117', '20161231']) AS d;

您可以在文档中阅读有关格式字符串的更多信息。 date和timestamp部分分别适用于PARSE_DATE和PARSE_TIMESTAMP。

编辑：根据您更新的问题，您可以将查询更改为以下内容：

WITH allTables as (
  SELECT
      event.date as date,
      count(*) as totalSessions,
      count(DISTINCT user_dim.app_info.app_instance_id) as uniqueUsers
  FROM `namehiddenonlyhere.*`
  CROSS JOIN
      UNNEST(event_dim) AS event
  WHERE
      event.name = 'session_start'
  AND 
      event.date <= FORMAT_DATE('%Y%m%d', DATE_SUB(CURRENT_DATE, INTERVAL 1 DAY))
  AND
      event.date >= FORMAT_DATE('%Y%m%d', DATE_SUB(CURRENT_DATE, INTERVAL 7 DAY))
  GROUP BY event.date

  UNION ALL 
  SELECT
      event.date as date,
      count(*) as totalSessions,
      count(DISTINCT user_dim.app_info.app_instance_id) as uniqueUsers
  FROM `namehiddenonlyhere.*`
  CROSS JOIN
      UNNEST(event_dim) AS event
  WHERE
      event.name = 'session_start'
  AND 
      event.date <= FORMAT_DATE('%Y%m%d', DATE_SUB(CURRENT_DATE, INTERVAL 1 DAY))
  AND
      event.date >= FORMAT_DATE('%Y%m%d', DATE_SUB(CURRENT_DATE, INTERVAL 7 DAY))
  GROUP BY event.date
) SELECT 'all apps' as target, PARSE_DATE('%Y%m%d', date) as datapoint_time, totalSessions datapoint_value FROM allTables;

我做的唯一修改就是改变：

date AS datapoint_time

为：

PARSE_DATE('%Y%m%d', date) as datapoint_time

Answer 2

您是否尝试使用此表达式？

SELECT CONVERT(DATETIME,'20170118')

SQL将yyyymmdd转换为BigQuery中的时间戳？

2 个答案: