Question

我有这个查询，我得到一些列的计数我的问题是计算所有这些计数结果在一个新的领域

SELECT 
CONCAT(u.firstname,' ',u.lastname) AS 'agent',
COUNT(CASE WHEN pa.answer_text LIKE '%yes%' THEN pa.answer_id END) AS yes,
COUNT(CASE WHEN pa.answer_text LIKE '%no%' THEN pa.answer_id END) AS NO,
COUNT(CASE WHEN l2u.lead_status_id IN (4,5,8,39) THEN l2u.lead_id END) AS pending,
COUNT(CASE WHEN l2u.lead_status_id IN (7,14,43) THEN l2u.lead_id END) AS wrong_number,

这是问题..这个的正确语法是什么 COUNT（待定+错误编号+是+否）AS'总计'

FROM USER u, poll_votes pv, poll_answers pa, lead_to_user_original l2u

WHERE u.user_id = pv.user_id
AND pv.answer_id = pa.answer_id
AND l2u.lead_id = pv.vote_lead_id
AND (pa.answer_client_one = '869' OR pa.answer_client_two = '869')
AND pv.vote_date BETWEEN '2013-07-01 00:00:01' AND '2013-07-17 23:59:59'
GROUP BY u.user_id

Answer 1

如何将它包装在子查询中？

SELECT  agent,  yes, no, pending, wrong_number,
        yes + no + pending + wrong_number AS Total
FROM
        (
            SELECT  CONCAT(u.firstname,' ',u.lastname) AS 'agent',
                    COUNT(CASE WHEN pa.answer_text LIKE '%yes%' THEN pa.answer_id END) AS yes,
                    COUNT(CASE WHEN pa.answer_text LIKE '%no%' THEN pa.answer_id END) AS NO,
                    COUNT(CASE WHEN l2u.lead_status_id IN (4,5,8,39) THEN l2u.lead_id END) AS pending,
                    COUNT(CASE WHEN l2u.lead_status_id IN (7,14,43) THEN l2u.lead_id END) AS wrong_number
            FROM    USER u
                    INNER JOIN poll_votes pv 
                        ON u.user_id = pv.user_id 
                    INNER JOIN poll_answers pa 
                        ON pv.answer_id = pa.answer_id
                    INNER JOIN lead_to_user_original l2u 
                        ON l2u.lead_id = pv.vote_lead_id

            WHERE   (pa.answer_client_one = '869' OR pa.answer_client_two = '869')
                    AND pv.vote_date BETWEEN '2013-07-01 00:00:01' AND '2013-07-17 23:59:59'
            GROUP   BY u.user_id
        ) subA

Answer 2

您无法在select中的另一个字段中的SELECT子句中访问别名。您可以将其“推送”到子查询，并执行以下操作：

SELECT a.*, pending + wrong_number + yes + no AS 'total'
FROM (
    SELECT 
        CONCAT(u.firstname,' ',u.lastname) AS 'agent',
        COUNT(CASE WHEN pa.answer_text LIKE '%yes%' THEN pa.answer_id END) AS yes,
        COUNT(CASE WHEN pa.answer_text LIKE '%no%' THEN pa.answer_id END) AS NO,
        COUNT(CASE WHEN l2u.lead_status_id IN (4,5,8,39) THEN l2u.lead_id END) AS pending,
        COUNT(CASE WHEN l2u.lead_status_id IN (7,14,43) THEN l2u.lead_id END) AS wrong_number,
    FROM USER u
    JOIN poll_votes pv
      ON u.user_id = pv.user_id
    JOIN poll_answers pa
      ON pv.answer_id = pa.answer_id
    JOIN lead_to_user_original l2u
      ON l2u.lead_id = pv.vote_lead_id
    WHERE (pa.answer_client_one = '869' OR pa.answer_client_two = '869')
      AND pv.vote_date BETWEEN '2013-07-01 00:00:01' AND '2013-07-17 23:59:59'
    GROUP BY u.user_id
) AS a

我还将您的查询更改为使用"ANSI-92 style" joins，我认为它更具可读性......

求和查询结果

2 个答案: