Question

我正在尝试创建Google Analytics（分析）数据的摘要：按小时和按来源属性（在我的情况下为服务）分类的会话，交易和转化率。我正在从汇总属性数据集中查询名为“ ga_realtime_view”的表。这是从“ ga_realtime_sessions_”表创建的虚拟视图，可让我们使用标准SQL。

为了具有服务列，我必须使用UNNEST操作。但是，当我这样做时，它将复制所有会话和事务值。

以下是查询：

SELECT
EXTRACT(HOUR FROM TIMESTAMP_SECONDS(visitStartTime) AT TIME ZONE 'Europe/Paris') AS Hour,
hits.sourcePropertyInfo.sourcePropertyDisplayName AS service,
IFNULL(SUM(totals.visits),0) as sessions,
IFNULL(SUM(totals.transactions),0) as transactions,
IFNULL(ROUND((SUM(totals.transactions)/SUM(totals.visits))*100,2),0) AS conversionRate
FROM `XX.ga_realtime_view` AS session, UNNEST(session.hits) AS hits
GROUP BY
Hour,
service
ORDER BY
Hour

我知道关于此重复问题还有其他主题，但是我似乎找不到适合我的情况的解决方案。

感谢您的帮助

Answer 1

听起来好像您要避免将表与数组连接（通过, UNNEST(session.hits)），因为这样做会导致所有总数重复。不过，尚不清楚您想对服务名进行什么处理：如果有多个服务名，您是否要返回所有这些服务的数组？这是一次方法：

SELECT
EXTRACT(HOUR FROM TIMESTAMP_SECONDS(visitStartTime) AT TIME ZONE 'Europe/Paris') AS Hour,
ARRAY(
   SELECT sourcePropertyInfo.sourcePropertyDisplayName
   FROM UNNEST(session.hits) AS hits
) AS service,
IFNULL(SUM(totals.visits),0) as sessions,
IFNULL(SUM(totals.transactions),0) as transactions,
IFNULL(ROUND((SUM(totals.transactions)/SUM(totals.visits))*100,2),0) AS conversionRate
FROM `XX.ga_realtime_view` AS session
GROUP BY
Hour,
service
ORDER BY
Hour

不过，您会注意到，查询现在给出一个错误，您无法按数组分组。如果您希望在匹配中仅提供一种服务，则可以仅提取其中一种：

SELECT
EXTRACT(HOUR FROM TIMESTAMP_SECONDS(visitStartTime) AT TIME ZONE 'Europe/Paris') AS Hour,
(SELECT MAX(sourcePropertyInfo.sourcePropertyDisplayName)
 FROM UNNEST(session.hits) AS hits) AS service,
IFNULL(SUM(totals.visits),0) as sessions,
IFNULL(SUM(totals.transactions),0) as transactions,
IFNULL(ROUND((SUM(totals.transactions)/SUM(totals.visits))*100,2),0) AS conversionRate
FROM `XX.ga_realtime_view` AS session
GROUP BY
Hour,
service
ORDER BY
Hour

BigQuery UNNEST重复值

1 个答案: