Sql server SQL SERVER 2012，1000万行，匹配同一表中的多个字段以获得结果

在名为

#Temp

的临时表中有1000万条记录。这些列是

PrimaryKey, FIELD1, FIELD2, FIELDS3, FIELD4, FIELD5, FIELD6, DateField

其中PrimaryKey为数值型，而除DateField外的所有其他字段均为VarChar

现在我想从这个表中获取格式为的重复记录

IsDuplicate, PrimaryKey1, PrimaryKey2, DateField1, DateField2

目前，以下查询正在运行，但执行起来需要很长时间

   With DATA
(SELECT d.FIELD1 AS key1, d.* FROM #TEMP d WHERE  d.FIELD1 IS NOT NULL
UNION ALL
SELECT d.FIELD2 AS key1, d.* FROM #TEMP d WHERE  d.FIELD2 IS NOT NULL
UNION ALL
SELECT d.FIELD3 AS key1, d.* FROM #TEMP d WHERE  d.FIELD3 IS NOT NULL
UNION ALL
SELECT d.FIELD4 AS key1, d.* FROM #TEMP d WHERE  d.FIELD4 IS NOT NULL
UNION ALL
SELECT d.FIELD5 AS key1, d.* FROM #TEMP d WHERE  d.FIELD5 IS NOT NULL
)
SELECT distinct IsDuplicate, PK1, PK2, DATE1, DATE2
FROM
(
SELECT CASE WHEN CASE WHEN T1.FIELD1 = T2.FIELD1 AND T1.FIELD1 IS NOT NULL AND T1.FIELD1 != ' ' THEN 1 ELSE 0 END+CASE
WHEN T1.FIELD2 = T2.FIELD2 AND T1.FIELD2 IS NOT NULL AND T1.FIELD2 != '' THEN 1 ELSE 0 END+CASE
WHEN T1.FIELD3 = T2.FIELD3 AND T1.FIELD3 IS NOT NULL AND T1.FIELD3 != '' THEN 1 ELSE 0 END+
CASE WHEN T1.FIELD4 = T2.FIELD4 AND T1.FIELD4 IS NOT NULL THEN 1 ELSE 0 END+
CASE WHEN T1.FIELD5 = T2.FIELD5 AND T1.FIELD5 IS NOT NULL AND T1.FIELD5 != '' THEN 1 ELSE 0 END+
CASE WHEN T1.FIELD6 = T2.FIELD6 AND T1.FIELD6 IS NOT NULL AND T1.FIELD6 != '' THEN 1 ELSE 0 END >= 4 THEN 1 ELSE 0 END AS IsDuplicate,
CASE
WHEN T1.FIELD1 = T2.FIELD1 AND T1.FIELD1 IS NOT NULL AND T1.FIELD1 != ' ' THEN 1 ELSE 0 END+
CASE WHEN T1.FIELD3 = T2.FIELD3 AND T1.FIELD3 IS NOT NULL AND T1.FIELD3 != '' THEN 1 ELSE 0 END+
CASE WHEN T1.FIELD4 = T2.FIELD4 AND T1.FIELD4 IS NOT NULL THEN 1 ELSE 0 END+
CASE WHEN T1.FIELD5 = T2.FIELD5 AND T1.FIELD5 IS NOT NULL AND T1.FIELD5 != '' THEN 1 ELSE 0 END+
CASE WHEN T1.FIELD6 = T2.FIELD6 AND T1.FIELD6 IS NOT NULL AND T1.FIELD6 != '' THEN 1 ELSE 0 END AS IsPotential, 
T1.PK AS PK1, T2.PK AS PK2, T1.Date AS DATE1, T2.Date AS DATE2
FROM Data T1
INNER JOIN Data T2 ON T1.key1 = T2.key1
) R where R.IsDuplicate = 1
OR R.IsPotential >= 3

寻找一种替代方法，以优化的方式获得所需的结果

---

谢谢。

索引？执行计划？没有相关信息就无法提供帮助？如果您有10万行，为什么要使用临时表？您是否向该表添加了索引？最后，期望的结果是什么？这个查询是做什么的？看起来它只是在尝试加载所有内容？@Alcapoon第二组案例中没有字段2是否是一个错误销售错误？是否有可能出现两次以上的值？您的输出格式可能无法很好地处理这种情况…您好！上面查询中使用的字段名是伪字段名…实际字段名类似于名字、姓氏、州、邮政编码、，电子邮件等……1000万行是从实际物理表加载到临时表中的相关字段，用于处理所需结果……在您的

案例
语句中，您执行的
T1.FIELD2=T2.FIELD2和T1.FIELD2不为NULL
，但根据
ANSI_NULLS
设置，通常不需要包含
T1.FIELD2不为NULL
，因为
NULL
不等于任何内容，包括另一个
NULL
：