Question

我正在使用下面的脚本（来自上一个问题）查询AWS Redshift，但排除一些referrer_url时遇到了麻烦。

例如，其中一些结构为https://google.com/erfkhjg/facebook/sdfdfd或https://bing.com/erfkhjg/facebook/sdfdfd，应计为search，但应计为social。

如何从social计数中排除这些链接，而将其计为search？我试图在第一个AND ... NOT LIKE块中添加多个SUM语句，但是没有用。感谢您的帮助！

SELECT A.page_id, A.views,
       SUM(CASE WHEN referrer_url LIKE '%%facebook%%' OR
                     referrer_url LIKE '%%instagram%%' OR
                     referrer_url LIKE '%%twitter%%'
                THEN 1 ELSE 0
           END) AS social,
       SUM(CASE WHEN referrer_url LIKE '%%google%%' OR
                     referrer_url LIKE '%%bing%%' OR
                     referrer_url LIKE '%%yahoo%%' 
                THEN 1 ELSE 0
           END) AS search
FROM table1 A LEFT JOIN
     table2 B
     ON B.page_id = A.page_id   
WHERE B.dt BETWEEN '20190401' AND '20190430'
GROUP BY A.page_id, A.views;

Answer 1

我认为，至少在您的示例中，查找域名前后的时间段是足够的：

SELECT A.page_id, A.views,
       SUM(CASE WHEN referrer_url LIKE '%.facebook.%' OR
                     referrer_url LIKE '%.instagram.%' OR
                     referrer_url LIKE '%.twitter.%'
                THEN 1 ELSE 0
           END) AS social,
       SUM(CASE WHEN referrer_url LIKE '%.google.%' OR
                     referrer_url LIKE '%.bing.%' OR
                     referrer_url LIKE '%.yahoo.%' 
                THEN 1 ELSE 0
           END) AS search
FROM table1 A LEFT JOIN
     table2 B
     ON B.page_id = A.page_id   
WHERE B.dt BETWEEN '20190401' AND '20190430'
GROUP BY A.page_id, A.views;

请注意，搜索模式中不需要两个通配符。

在WHEN语句中具有多个LIKE和NOT LIKE的SQL查询

1 个答案: