计算 bigquery 中的唯一网页浏览量

Question

我正在尝试计算这 2 个页面的唯一网页浏览量。

• 注册帐户 ->mysite.com/form?account=true&subscribed
• 仅限帐户 -> mysite.com/form?account=true

我有这个带有 case 函数的查询，但是当我为每个页面参数单独运行以下查询时，我得到的结果与运行它组合时不同。但是综合结果是不准确的。有人可以告诉我我在这里做错了什么吗？

     SELECT
    COUNT(DISTINCT (CASE WHEN hits.type = "PAGE" THEN CONCAT(fullvisitorid, 
    STRING(visitid), hits.page.pagepath) END)) AS UniquePageViews,
    CASE WHEN (REGEXP_MATCH (hits.page.pagePath, '(.*account=true)'))  THEN "Accounts" 
    WHEN (REGEXP_MATCH (hits.page.pagePath, '(.*subscribed)')) THEN "Signups" ELSE "Others" END AS Goals
    FROM
     [mydata.ga_sessions_20150506]
    GROUP BY
     Goals

Answer 1

问题可能是字符串"account=true" 和"subscribed" 出现在同一个pagePath 中的情况，因此当您尝试计算它们时，只考虑一个。

解决方法之一是改变匹配条件，如：

SELECT
    EXACT_COUNT_DISTINCT(CASE WHEN hits.type = "PAGE" THEN CONCAT(fullvisitorid, 
STRING(visitid), hits.page.pagepath) END) AS UniquePageViews,
    CASE WHEN (REGEXP_MATCH (hits.page.pagePath, '(account=true)') AND NOT REGEXP_MATCH (hits.page.pagePath, '(subscribed)'))  THEN "Accounts" 
         WHEN (REGEXP_MATCH (hits.page.pagePath, '(subscribed)') AND NOT REGEXP_MATCH (hits.page.pagePath, '(account=true)')) THEN "Signups"
         WHEN (REGEXP_MATCH (hits.page.pagePath, '(subscribed)') AND  REGEXP_MATCH (hits.page.pagePath, '(account=true)')) THEN "Both"
         ELSE "Others" END AS Goals
FROM
 [mydata.ga_sessions_20150506]
GROUP BY
 Goals

我强制条件仅匹配 "account=true" 而不是 "subscribed"。

例如，这是我在 ga_sessions 数据集上测试的：

SELECT
exact_count_distinct(CASE WHEN hits.type = "PAGE" THEN CONCAT(fullvisitorid, STRING(visitid), hits.page.pagepath) END) AS UniquePageViews,
CASE WHEN (REGEXP_MATCH (hits.page.pagePath, '(colcci)') AND NOT REGEXP_MATCH (hits.page.pagePath, '(lacoste)'))  THEN "colcci" 
     WHEN (REGEXP_MATCH (hits.page.pagePath, '(lacoste)') AND NOT REGEXP_MATCH (hits.page.pagePath, '(colcci)')) THEN "lacoste"
     WHEN (REGEXP_MATCH (hits.page.pagePath, '(lacoste)') AND REGEXP_MATCH (hits.page.pagePath, '(colcci)')) THEN 'both'
     ELSE "Others" END AS Goals
FROM [40663402.ga_sessions_20150506]
GROUP BY
Goals

希望这会有所帮助。有任何问题，请告诉我们，