C# sql查询中的多选择命令

我想在我的sql中检索此类表：

    Type     Qad (type count)     Correct(No. of item without error)     Accuracy(%)
   type 1        3                                 3                            100
   type2         6                                 3                             50

我正在尝试使用此查询，但它有一个错误：

SELECT `type` AS 'ReturnType', COUNT(*) AS 'QAd',
(SELECT COUNT(*) FROM bartran WHERE QASample='2' AND QAErrorID='1' AND `type`=ReturnType AND TimefileDate BETWEEN '2012-01-01' AND '2012-12-31' AS 'Correct', 
(SELECT COUNT(*) FROM bartran WHERE QASample='2' AND QAErrorID='1' AND `type`=ReturnType AND TimefileDate BETWEEN '2012-01-01' AND '2012-12-31') / COUNT(*) * 100) AS 'Accuracy' 
FROM bartran WHERE QASample='2'
AND TimefileDate BETWEEN '2012-01-01' AND '2012-12-31'
GROUP BY TYPE;

---

我现在是新来的，希望有人能帮我。谢谢

查询中缺少括号），正确答案是：

SELECT `type`  AS 'ReturnType', 
       Count(*)   AS 'QAd', 
       (SELECT Count(*) 
        FROM   bartran 
        WHERE  qasample = '2' 
               AND qaerrorid = '1' 
               AND `type` = returntype 
               AND timefiledate BETWEEN '2012-01-01' AND '2012-12-31') AS 
       'Correct', 
       (SELECT Count(*) 
        FROM   bartran 
        WHERE  qasample = '2' 
               AND qaerrorid = '1' 
               AND `type` = returntype 
               AND timefiledate BETWEEN '2012-01-01' AND '2012-12-31') / Count(* 
       ) * 100 AS 'Accuracy' 
FROM   bartran 
WHERE  qasample = '2' 
       AND timefiledate BETWEEN '2012-01-01' AND '2012-12-31' 
GROUP  BY type; 

更好的查询是：

SELECT `type` AS 'ReturnType', 
       Count(*)  AS 'QAd', 
       Count(CASE 
               WHEN qaerrorid = '1' THEN 1 
               ELSE NULL 
             end)  AS 'Correct', 
       Count(CASE 
               WHEN qaerrorid = '1' THEN 1 
               ELSE NULL 
             end) / Count(*) * 100 AS 'Accuracy' 
FROM   bartran 
WHERE  qasample = '2' 
       AND timefiledate BETWEEN '2012-01-01' AND '2012-12-31' 
GROUP  BY type;

---

查询中缺少括号），正确答案是：

SELECT `type`  AS 'ReturnType', 
       Count(*)   AS 'QAd', 
       (SELECT Count(*) 
        FROM   bartran 
        WHERE  qasample = '2' 
               AND qaerrorid = '1' 
               AND `type` = returntype 
               AND timefiledate BETWEEN '2012-01-01' AND '2012-12-31') AS 
       'Correct', 
       (SELECT Count(*) 
        FROM   bartran 
        WHERE  qasample = '2' 
               AND qaerrorid = '1' 
               AND `type` = returntype 
               AND timefiledate BETWEEN '2012-01-01' AND '2012-12-31') / Count(* 
       ) * 100 AS 'Accuracy' 
FROM   bartran 
WHERE  qasample = '2' 
       AND timefiledate BETWEEN '2012-01-01' AND '2012-12-31' 
GROUP  BY type; 

更好的查询是：

SELECT `type` AS 'ReturnType', 
       Count(*)  AS 'QAd', 
       Count(CASE 
               WHEN qaerrorid = '1' THEN 1 
               ELSE NULL 
             end)  AS 'Correct', 
       Count(CASE 
               WHEN qaerrorid = '1' THEN 1 
               ELSE NULL 
             end) / Count(*) * 100 AS 'Accuracy' 
FROM   bartran 
WHERE  qasample = '2' 
       AND timefiledate BETWEEN '2012-01-01' AND '2012-12-31' 
GROUP  BY type;

---

声明返回的错误。从bartran中选择Type、Qad、Correct、accurity，其中QASample='2'，以及其他您希望数据类似于分组类型的内容；声明返回的错误。从bartran中选择Type、Qad、Correct、accurity，其中QASample='2'，以及其他您希望数据类似于分组类型的内容；我认为这解决了我的问题，但为什么执行要花很多时间呢？1分钟以上：（@JRC尝试我的新查询added@Ankur最好使用SUM而不是Count

SUM（当QARRORID='1'然后1 ELSE 0结束时的情况）

。也可以使用

Count（1）

，而不是

Count（*）

。这会有点麻烦faster@AndreyVoloshin计数（1）和计数（*）都是一样的，请看，我也没有发现sum是否优于count，所以我更喜欢count，因为代码清晰。我认为这现在解决了我的问题，但为什么执行会花费很多时间？1分钟多一点。：（@JRC尝试我的新查询added@Ankur使用SUM代替Count

SUM是一个很好的方法（当QARRORID='1'然后1 ELSE 0结束时的情况）

。也使用

COUNT（1）

，而不是

COUNT（*）

。这将是一个小错误faster@AndreyVoloshin计数（1）和计数（*）都是一样的，请看，而且我也没有发现sum是否比count好，所以从代码清晰的角度来看，我更喜欢count