在计算列上使用聚合函数

Question

尝试在新创建的计算列上使用sum（）。 求和函数无法识别calc列名称。 这是我的代码

select name, is_open,    
       CASE WHEN text like '%perfect%' or '%amazing%' or '%happy%' or '%delicious%'or '%fabulous%'or '%fantastic%'or '%kind%'   THEN 1
            WHEN text like '%hate%'or '%horrible%'or '%bad%'or '%angry%'or '%fantastic%'or '%expensive%'or '%disgusting%' THEN -1
       END sentiment_rating,    
       sum(sentiment_rating) as sum
from review as r left join
     business as b
     on b.id = r.business_id
where is_open is not Null and sentiment_rating is not Null
group by name
order by name ASC

Answer 1

同时选择未聚合的列和聚合的列都没有意义。 因此，只求sum()表达式为：

select name, is_open,    
       sum(CASE WHEN text like '%perfect%' or text like '%amazing%' or text like '%happy%' or text like '%delicious%' or text like '%fabulous%' or text like '%fantastic%' or text like '%kind%'   THEN 1
                WHEN text like '%hate%' or text like '%horrible%' or text like '%bad%' or text like '%angry%' or text like '%fantastic%'or '%expensive%' or text like '%disgusting%' THEN -1
           END) as sum_sentiment_rating 
from review as r left join
     business as b
     on b.id = r.business_id
where is_open is not Null and sentiment_rating is not Null
group by name, is_open
order by name ASC;

我不确定是要每个name一行还是每个name / is_open一行。 我假设是后者，然后将is_open添加到GROUP BY 。

Answer 2

您所有的“ LIKE条件在语法上都是错误的。
您不能拥有：

text like '%perfect%' or '%amazing%'....

正确的语法是：

text like '%perfect%' or text like '%amazing%' ....

然后，您必须直接对使用CASE计算的列求和，并将is_open包含在分组列中：

select name, is_open,    
  SUM(CASE 
    WHEN text like '%perfect%' or text like '%amazing%' or text like '%happy%' or text like '%delicious%' or text like '%fabulous%' or text like '%fantastic%' or text like '%kind%' THEN 1
    WHEN text like '%hate%' or text like '%horrible%' or text like '%bad%' or text like '%angry%' or text like '%fantastic%' or text like '%expensive%' or text like '%disgusting%' THEN -1
  END) as sumrating
from review as r left join
     business as b
     on b.id = r.business_id
where is_open is not Null and sentiment_rating is not Null
group by name, is_open
order by name ASC