1. query optimization, should try to avoid full table scan, should first consider indexing by the column involved in where and order.
2. fields should be avoided to a null value is determined in the where clause, will cause the engine to give up using the index and a full table scan, such as: select id from t where num is null may be provided on a default value of 0 num, to ensure num table column value is not null, then this query: select id from t where num = 0
3 should be avoided in the where clause! = Or <> operator, otherwise the engine will give up using the index and a full table scan.
4 should be avoided in the where clause to connect or condition will cause the engine to give up using the index and a full table scan, such as: select id from t where num = 10 or num = 20 can do a query: select id from t where num = 10 union all select id from t where num = 20
5.in should be used with caution and not in, otherwise it will lead to a full table scan, such as: select id from t where num in (1,2,3) for continuous values between can not use in the: select id from t where num between 1 and 3
6. The following query will also lead to a full table scan: select id from t where name like '% Li' To improve efficiency, can be considered full-text search.
7. If the parameters in the where clause, will lead to a full table scan. Because SQL at runtime will only resolve local variables, but the optimizer can not defer the choice of access plan to run; it must choose at compile time. If, however, establish access plan at compile time, the value of the variable is unknown, and therefore can not be used as an index entry selected. As the following statement will perform full table scan: select id from t where num = @ num can be changed to force the query using the index: select id from t with (index (index name)) where num = @ num
8. fields should be avoided to operate in the where clause expressions, which will cause the engine to give up using the index and a full table scan. Such as: select id from t where num / 2 = 100 should be changed to: select id from t where num = 100 * 2
9. The field should be avoided to a function operation in the where clause that will cause the engine to give up using the index and full table scan. Such as: select id from t where substring (name, 1,3) = 'abc', name beginning with the id abc
Should read:
select id from t where name like 'abc%'
10. Do not carry out functions, arithmetic operations, or other expressions in the where clause "=" left, or the system may not work properly indexed.
11. As a condition of using the index field, if the index is a composite index, you must use the index to the first field to ensure the system uses the index as a condition, otherwise the index will not be used, and should as much as possible so that field order is consistent with the order index.
12. Do not write the query does not make sense, such as the need to create an empty table structure: select col1, col2 into #t from t where 1 = 0
This code does not return any result sets, but consumes system resources and should be replaced by this:
create table #t(...)
13.很多时候用 exists 代替 in 是一个好的选择：select num from a where num in(select num from b)
用下面的语句替换：
select num from a where exists(select 1 from b where num=a.num)
14.并不是所有索引对查询都有效，SQL是根据表中数据来进行查询优化的，当索引列有大量数据重复时，SQL查询可能不会去利用索引，如一表中有字段sex，male、female几乎各一半，那么即使在sex上建了索引也对查询效率起不了作用。
15. 索引并不是越多越好，索引固然可 以提高相应的 select 的效率，但同时也降低了 insert 及 update 的效率，因为 insert 或 update 时有可能会重建索引，所以怎样建索引需要慎重考虑，视具体情况而定。一个表的索引数最好不要超过6个，若太多则应考虑一些不常使用到的列上建的索引是否有 必要。
16. 应尽可能的避免更新 clustered 索引数据列，因为 clustered 索引数据列的顺序就是表记录的物理存储顺序，一旦该列值改变将导致整个表记录的顺序的调整，会耗费相当大的资源。若应用系统需要频繁更新 clustered 索引数据列，那么需要考虑是否应将该索引建为 clustered 索引。
17.尽量使用数字型字段，若只含数值信息的字段尽量不要设计为字符型，这会降低查询和连接的性能，并会增加存储开销。这是因为引擎在处理查询和连接时会逐个比较字符串中每一个字符，而对于数字型而言只需要比较一次就够了。
18.尽可能的使用 varchar/nvarchar 代替 char/nchar ，因为首先变长字段存储空间小，可以节省存储空间，其次对于查询来说，在一个相对较小的字段内搜索效率显然要高些。
19.任何地方都不要使用 select * from t ，用具体的字段列表代替“*”，不要返回用不到的任何字段。
20.尽量使用表变量来代替临时表。如果表变量包含大量数据，请注意索引非常有限（只有主键索引）。
21.避免频繁创建和删除临时表，以减少系统表资源的消耗。
22.临时表并不是不可使用，适当地使用它们可以使某些例程更有效，例如，当需要重复引用大型表或常用表中的某个数据集时。但是，对于一次性事件，最好使用导出表。
23.在新建临时表时，如果一次性插入数据量很大，那么可以使用 select into 代替 create table，避免造成大量 log ，以提高速度；如果数据量不大，为了缓和系统表的资源，应先create table，然后insert。
24.如果使用到了临时表，在存储过程的最后务必将所有的临时表显式删除，先 truncate table ，然后 drop table ，这样可以避免系统表的较长时间锁定。
25.尽量避免使用游标，因为游标的效率较差，如果游标操作的数据超过1万行，那么就应该考虑改写。
26.使用基于游标的方法或临时表方法之前，应先寻找基于集的解决方案来解决问题，基于集的方法通常更有效。
27. 与临时表一样，游标并不是不可使 用。对小型数据集使用 FAST_FORWARD 游标通常要优于其他逐行处理方法，尤其是在必须引用几个表才能获得所需的数据时。在结果集中包括“合计”的例程通常要比使用游标执行的速度快。如果开发时 间允许，基于游标的方法和基于集的方法都可以尝试一下，看哪一种方法的效果更好。
28.在所有的存储过程和触发器的开始处设置 SET NOCOUNT ON ，在结束时设置 SET NOCOUNT OFF 。无需在执行存储过程和触发器的每个语句后向客户端发送DONE_IN_PROC 消息。
29.尽量避免大事务操作，提高系统并发能力。
30.尽量避免向客户端返回大数据量，若数据量过大，应该考虑相应需求是否合理。