数据库

NOT IN 与NULL的比较

肉丝 — Fri, 30 Mar 2007 15:07:00 GMT

在我们的语句中有时有用到NOT IN 语句，但是用这条语句时存在一定的问题。
1.NOT IN
SELECT '有' WHERE 1 NOT IN(20, NULL)

---
（所影响的行数为 0 行）

分析：按理说1本身就不是20, NULL值里，应该有返回值。
2.使用IN
SELECT '有' WHERE 1 IN(1, NULL)

---
有
（所影响的行数为 1 行）
3.根据SQL SERVER的说明，被比较值的类型比较与比较值的类型一值， NULL不算任何类型就像VARIANT中的VT_NULL一样只一个标志吧。
例如：
SELECT '有' WHERE 1 NOT IN(20, CAST(NULL AS INT))
---
（所影响的行数为 0 行）
结果还是为0。

总结：在比较时将被比较值转换为比较值的类型，所以我们用时要注意。其次在比较值大小等操作时如无特殊需求就将为NULL的去掉。
例： SELECT * FROM TABLE_NAME WHERE COLUMN_NAME IS NULL
--说明不能用 COLUMN_NAME = NULL哦。相反的就是NOT COLUMN_NAME IS NULL 或者 COLUMN_NAME IS NOT NULL
为啥原因不行，上面我已经讲了哈。

用ODBC方式读取EXCEL时，读取出来的内容无效（#错误)之类的解决方法。

肉丝 — Wed, 28 Mar 2007 03:43:00 GMT

在ACCESS中会出现#数字!, 而在程序或SQL中会出现NULL

示例数据
字段1 字段2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
吕
ii
ddd

1.如果用SELECT * FROM [EXCEL 5.0;DATABASE=C:\Test.xls].[sheet1$]查询出来，那么 22后面的全为NULL值
2.那么用SELECT * FROM [EXCEL 5.0;HDR=YES;IMEX=1;DATABASE=C:\Test.Xls].[sheet1$]查询，后面的值还会是NULL，为什么呢？在这种情况下前八行必须有一个是以符存储的，也就是在单元格前加一个单引号（'1)，用这种方法就OK了。
3.那么如果前8行没有用又咋办呢？
那么把第一行不当做字段名就行了
SELECT * FROM [EXCEL 5.0;HDR=NO;IMEX=1;DATABASE=C:\Test.xls].[sheet1]

好了，三种方案根据实际情况任选喽。

ADO连接字符串

肉丝 — Fri, 17 Jun 2005 03:12:00 GMT

Note:
这些内容是从codeproject转贴来的，偶英文不好只翻译了一部分，加之本人很懒，所以有什么不妥之处还请指证

 DSN-Less Connection

 DSN Connection

ODBC DSN

 OLE DB Provider

 OLE DB Data Link Connections

Data Link File - UDL

1.ODBC Driver for dBASE
strConnection = _T("Driver={Microsoft dBASE Driver (*.dbf)};DriverID=277;" "Dbq=c:\\DatabasePath;");
注释:你必须在查询语句中指定文件名,例如:
CString strQuery = _T("Select Name, Address From Clients.dbf");
2.ODBC Driver for Excel
strConnection = _T("Driver={Microsoft Excel Driver (*.xls)};DriverId=790;
dbq=C:\\DatabasePath\\DBSpreadSheet.xls;DefaultDir=c:\\databasepath;");

3.ODBC Driver for Text

strConnection = _T("Driver={Microsoft Text Driver (*.txt; *.csv)};"

"Dbq=C:\\DatabasePath\\;Extensions=asc,csv,tab,txt;");

如果你用带制表分隔符的文件，必须创建schema.ini文件,并且指明格式在连接字符串中(Format=制表符.
注释:必须在查询语句中指定文件名做为表名
CString strQuery = _T("Select Name, Address From Clients.csv");

4.Visual FoxPro

a.strConnection = _T("Driver={Microsoft Visual Foxpro Driver};UID=;
SourceType=DBC;SourceDB=C:\\DatabasePath\\MyDatabase.dbc;Exclusive=No");

b.strConnection = _T("Driver={Microsoft Visual Foxpro Driver};UID=;"

"SourceType=DBF;SourceDB=C:\\DatabasePath\\MyDatabase.dbc;Exclusive=No");

5.ODBC Driver for Access

strConnection = _T("Driver={Microsoft Access Driver (*.mdb)};"

"Dbq=c:\\DatabasePath\\dbaccess.mdb;Uid=;Pwd=;");

如果使用工作组文件，需在连接字符串中指定工作组文件路径，用户名及密码。
有两种方式，第一种是在连接串中说明，另一种是使用程序方式在打开时指明。
例如:
strConnection = _T("Driver={Microsoft Access Driver (*.mdb)};"
        "Dbq=C:\\VC Projects\\ADO\\Samples\\AdoTest\\dbTestSecurity.mdb;"
        "SystemDB=C:\\Program Files\\Microsoft Office\\Office\\SYSTEM.mdw;"
        "Uid=Carlos Antollini;Pwd=carlos");
或
strConnection = _T("Driver={Microsoft Access Driver (*.mdb)};"
        "Dbq=C:\\VC Projects\\ADO\\Samples\\AdoTest\\dbTestSecurity.mdb;"
        "SystemDB=C:\\Program Files\\Microsoft Office\\Office\\SYSTEM.mdw;");
if(pDB.Open(strConnection, "DatabaseUser", "DatabasePass"))
{
    DoSomething();
    pDB.Close();
}

以独占的方式
strConnection = _T("Driver={Microsoft Access Driver (*.mdb)};"

"Dbq=c:\\DatabasePath\dbaccess.mdb;Exclusive=1;");

6.ODBC Driver for SQL Server

a.SQL SERVER认证方式
strConnection = _T("Driver={SQL Server};Server=MyServerName;"
"Trusted_Connection=no;"
"Database=MyDatabaseName;Uid=MyUserName;Pwd=MyPassword;");

b.混合认证方式
strConnection = _T("Driver={SQL Server};Server=MyServerName;"

"Database=myDatabaseName;Uid=;Pwd=;");

WINDOWS NT认证方式
strConnection = _T("Driver={SQL Server};Server=MyServerName;"

"Database=MyDatabaseName;Trusted_Connection=yes;");

如果SQL SERVER运行在本地计算机上
strConnection = _T("Driver={SQL Server};Server=(local);"

"Database=MyDatabaseName;Uid=MyUserName;Pwd=MyPassword;");
如果计算机上存在两个实例只需在服务名上加“\实例名“
dbnmpntw Win32 Named Pipes dbmssocn Win32 Winsock TCP/IP

dbmsspxn Win32 SPX/IPX
dbmsvinn Win32 Banyan Vines
dbmsrpcn Win32 Multi-Protocol (Windows RPC)

更多信息请参见: Q238949.

strConnection = _T("Driver={SQL Server};Server=130.120.110.001;"

"Address=130.120.110.001,1052;Network=dbmssocn;Database=MyDatabaseName;"

"Uid=myUsername;Pwd=myPassword;");

8.Driver for Oracle

新的ODBC格式

strConnect = _T("Driver={Microsoft ODBC for Oracle};Server=OracleServer.world;"

"Uid=MyUsername;Pwd=MyPassword;");

旧的ODBC方式
strConnect = _T("Driver={Microsoft ODBC Driver for Oracle};"

"ConnectString=OracleServer.world;Uid=myUsername;Pwd=myPassword;");

9.ODBC Driver for MySQL

连接到本数据库:

strConnect = _T("Driver={MySQL ODBC 3.51 Driver};Server=localhost;"

"Database=MyDatabase;User=MyUserName;Password=MyPassword;Option=4;");

连接到远程数据库，需要在SERVER参数指定服务名或IP地址,如果端口不是默认端口3306,你必面指定。

strConnect = _T("Driver={mySQL ODBC 3.51 Driver};Server=MyRemoteHost;"

"Port=3306;Option=4;Database=MyDatabase;Uid=MyUsername;Pwd=MyPassword;");

The parameter Option can be one or more of the following values:

1 - The client can't handle that MyODBC returns the real width of a column.
2 - The client can't handle that MySQL returns the true value of affected rows. If this flag is set then MySQL returns 'found rows' instead. One must have MySQL 3.21.14 or newer to get this to work.
4 - Make a debug log in c:\myodbc.log. This is the same as putting MYSQL_DEBUG=d:t:O,c::\myodbc.log in AUTOEXEC.BAT.
8 - Don't set any packet limit for results and parameters.
16 - Don't prompt for questions even if driver would like to prompt.
32 - Enable or disable the dynamic cursor support. This is not allowed in MyODBC 2.50.
64 - Ignore use of database name in 'database.table.column'.
128 - Force use of ODBC manager cursors (experimental).
256 - Disable the use of extended fetch (experimental).
512 - Pad CHAR fields to full column length.
1024 - SQLDescribeCol() will return fully qualified column names.
2048 - Use the compressed server/client protocol.
4096 - Tell server to ignore space after function name and before '(' (needed by PowerBuilder). This will make all function names keywords!
8192 - Connect with named pipes to a MySQLd server running on NT.
16384 - Change LONGLONG columns to INT columns (some applications can't handle LONGLONG).
32768 - Return 'user' as Table_qualifier and Table_owner from SQLTables (experimental).
65536 - Read parameters from the client and ODBC groups from my.cnf.
131072 - Add some extra safety checks (should not be needed but...).

If you want to have multiple options, you should add the above flags! For example: 16 + 1024 = 1030 and use Option= 1030;.

For more information, go to MyODBC Reference Manual.

ODBC Driver for AS400

strConnect = _T("Driver={Client Access ODBC Driver (32-bit)};System=myAS400;"

"Uid=myUsername;Pwd=myPassword;");

ODBC Driver for SyBase

strConnect = _T("Driver={Sybase System 10};Srvr=MyServerName;Uid=MyUsername;"

"Pwd=myPassword;");

ODBC Driver for Sybase SQL AnyWhere

strConnect = _T("ODBC;Driver=Sybase SQL Anywhere 5.0;"

"DefaultDir=C:\\DatabasePath\;Dbf=C:\\SqlAnyWhere50\\MyDatabase.db;"

"Uid=MyUsername;Pwd=MyPassword;Dsn=\"\";");

DSN Connection

ODBC DSN

strConnect = _T("DSN=MyDSN;Uid=MyUsername;Pwd=MyPassword;");

OLE DB Provider

OLE DB Provider for SQL Server

For Standard security:

strConnect = _T("Provider=sqloledb;Data Source=MyServerName;"

"Initial Catalog=MyDatabaseName;"

"User Id=MyUsername;Password=MyPassword;");

For Trusted Connection security (Microsoft Windows NT integrated security):

strConnect = _T("Provider=sqloledb;Data Source=MyServerName;"

"Initial Catalog=MyDatabaseName;"

"Integrated Security=SSPI;");

If you want to connect to a "Named Instance" (SQL Server 2000), you must to specify Data Source=Servere Name\Instance Name like in the following example:

strConnect = _T("Provider=sqloledb;Data Source=MyServerName\MyInstanceName;"

"Initial Catalog=MyDatabaseName;User Id=MyUsername;Password=MyPassword;");

If you want to connect with a SQL Server running on the same computer, you must specify the keyword (local) in the Data Source like in the following example:

strConnect = _T("Provider=sqloledb;Data Source=(local);"

"Initial Catalog=myDatabaseName;"

"User ID=myUsername;Password=myPassword;");

To connect to SQL Server running on a remote computer (via an IP address):

strConnect = _T("Provider=sqloledb;Network Library=DBMSSOCN;"

"Data Source=130.120.110.001,1433;"

"Initial Catalog=MyDatabaseName;User ID=MyUsername;"

"Password=MyPassword;");

OLE DB Provider for MySQL (By Todd Smith)

strConnection = _T("Provider=MySQLProv;Data Source=test");

Where test is the name of MySQL database. Also, you can replace the name of the database by the following connection string: server=localhost;DB=test.

OLE DB Provider for AS400

strConnect = _T("Provider=IBMDA400;Data source=myAS400;User Id=myUsername;"

"Password=myPassword;");

更多信息请参见: Using the OLE DB Provider for AS/400 and VSAM.

OLE DB Provider for Active Directory

strConnect = _T("Provider=ADSDSOObject;User Id=myUsername;Password=myPassword;");

OLE DB Provider for DB2

If you are using a TCP/IP connection:

strConnect = _T("Provider=DB2OLEDB;Network Transport Library=TCPIP;"

"Network Address=130.120.110.001;"

"Initial Catalog=MyCatalog;Package Collection=MyPackageCollection;"

"Default Schema=MySchema;User ID=MyUsername;Password=MyPassword;");

If you are using APPC connection:

strConnect = _T("Provider=DB2OLEDB;APPC Local LU Alias=MyLocalLUAlias;"

"APPC Remote LU Alias=MyRemoteLUAlias;Initial Catalog=MyCatalog;"

"Package Collection=MyPackageCollection;Default Schema=MySchema;"

"User ID=MyUsername;Password=MyPassword;");

更多信息请参见: Using the OLE DB Provider for DB2.

OLE DB Provider for Microsoft Jet

Connecting to an Access file using the JET OLE DB Provider:

Using Standard security:

strConnect = _T("Provider=Microsoft.Jet.OLEDB.4.0;"

"Data Source=C:\\DatabasePath\\MmDatabase.mdb;"

"User Id=admin;Password=;");

If you are using a Workgroup (System database):

strConnect = _T("Provider=Microsoft.Jet.OLEDB.4.0;"

"Data Source=C:\\DataBasePath\\mydb.mdb;"

"Jet OLEDB:System Database=MySystem.mdw;");

pRs.Open(strConnect, "MyUsername", "MyPassword");

Connecting to an Excel Spreadsheet using the JET OLE DB Provider:

· strConnect = _T("Provider=Microsoft.Jet.OLEDB.4.0;"

· "Data Source=C:\\DatabasePath\\DBSpreadSheet.xls;"

"Extended Properties=\"\"Excel 8.0;HDR=Yes;\"\";");

Note: If "HDR=Yes", the provider will not include the first row of the selection into the recordset. If "HDR=No", the provider will include the first row of the cell range (or named ranged) into the recordset.

更多信息请参见: Q278973.

Connecting to a Text file using the JET OLE DB Provider:

· strConnect =

· _T("Provider=Microsoft.Jet.OLEDB.4.0;Data Source=C:\\DatabasePath\\;"

· "Extended Properties=\"\"text;"

"HDR=Yes;FMT=Delimited;\"\";");

Note: You must specify the filename in the SQL statement... For example:

CString strQuery = _T("Select Name, Address From Clients.txt");

更多信息请参见: Q262537.

Connecting to an Outlook 2000 personal mail box using the JET OLE DB Provider: (By J. Cardinal)

· strConnect = _T("Provider=Microsoft.Jet.OLEDB.4.0;Outlook 9.0;"

"MAPILEVEL=;DATABASE=C:\\Temp\\;")

Replace c:\temp with any temporary folder. It will create a schema file in that folder when you open it which shows all the fields available. Blank MAPILEVEL indicates top level of folders).

Connecting to an Exchange mail box through JET: (By J. Cardinal)

· strConnect = _T("Provider=Microsoft.Jet.OLEDB.4.0;Exchange 4.0;"

"MAPILEVEL=Mailbox - Pat Smith|;DATABASE=C:\\Temp\\;")

You must replace c:\temp with any temporary folder.

Replace Pat Smith with the name of the mail box and you must keep vertical pipe character | to indicate top level of folders. Place sub folder after vertical pipe if accessing specific folder.

Note: you can enter queries against the mail store just like a database... For example:

CString strQuery = _T("SQL "SELECT Contacts.* FROM Contacts;");

更多信息请参见: The Jet 4.0 Exchange/Outlook IISAM.

OLE DB Provider for ODBC Databases

If you want to connect with a Microsoft Access database:

strConnect = _T("Provider=MSDASQL;Driver={Microsoft Access Driver (*.mdb)};"

"Dbq=c:\\DatabasePath\\MyDatabase.mdb;Uid=MyUsername;Pwd=MyPassword;");

If you want to connect with a SQL Server database:

strConnect = _T("Provider=MSDASQL;Driver={SQL Server};Server=MyServerName;"

"Database=MyDatabaseName;Uid=MyUsername;Pwd=MyPassword;");

If you want to use DSN:

strConnect = _T("Provider=MSDASQL;PersistSecurityInfo=False;"

"Trusted_Connection=Yes;"

"Data Source=MyDSN;catalog=MyDatabase;");

更多信息请参见: Microsoft OLE DB Provider for ODBC.

OLE DB Provider for OLAP

Microsoft® OLE DB for Online Analytical Processing (OLAP) is a set of objects and interfaces that extends the ability of OLE DB to provide access to multidimensional data stores.

strConnect = _T("Provider=MSOLAP;Data Source=MyOLAPServerName;"

"Initial Catalog=MyOLAPDatabaseName;");

Connection using HTTP:

This feature enables a client application to connect to an Analysis server through Microsoft® Internet Information Services (IIS) by specifying a URL in the Data Source property in the client application's connection string. This connection method allows PivotTable® Service to tunnel through firewalls or proxy servers to the Analysis server. A special Active Server Pages (ASP) page, Msolap.asp, enables the connection through IIS. The directory in which this file resides must be included as part of the URL when connecting to the server (for example, http://www.myserver.com/myolap/).

Using a URL

strConnect = _T("Provider=MSOLAP;Data Source=http://MyOLAPServerName/;"

"Initial Catalog=MyOLAPDatabaseName;");

Using SSL

strConnect = _T("Provider=MSOLAP;Data Source=https://MyOLAPServerName/;"

"Initial Catalog=MyOLAPDatabaseName;");

更多信息请参见: OLE DB for OLAP, Connecting Using HTTP.

OLE DB Provider for Oracle

OLE DB Provider for Oracle (from Microsoft)

The Microsoft OLE DB Provider for Oracle allows ADO to access Oracle databases.

strConnect = _T("Provider=MSDAORA;Data Source=MyOracleDB;User Id=myUsername;"

"Password=myPassword;");

更多信息请参见: Microsoft OLE DB Provider for Oracle.

OLE DB Provider for Oracle (from Oracle).

For Standard security:

strConnect = _T("Provider=OraOLEDB.Oracle;Data Source=MyOracleDB;"

"User Id=myUsername;Password=myPassword;");

For a Trusted connection:

OS Authenticated connect setting user ID to "/":

· strConnect = _T("Provider=OraOLEDB.Oracle;Data Source=MyOracleDB;"

"User Id=/;Password=;");

OS Authenticated connect using OSAuthent:

strConnect = _T("Provider=OraOLEDB.Oracle;Data Source=MyOracleDB;OSAuthent=1;")

Note: "Data Source=" must be set to the appropriate Net8 name which is known to the naming method in use. For example, for Local Naming, it is the alias in the tnsnames.ora file; for Oracle Names, it is the Net8 Service Name.

更多信息请参见: Oracle Provider for OLE DB Developer's Guide.

OLE DB Provider for Visual FoxPro

strConnect = _T("Provider=vfpoledb;"

"Data Source=C:\\DatabasePath\\MyDatabase.dbc;");

更多信息请参见: Microsoft OLE DB Provider for Visual FoxPro.

OLE DB Provider for Index Server (By Chris Maunder)

strConnect = _T("provider=msidxs;Data Source=MyCatalog;");

更多信息请参见:: Microsoft OLE DB Provider for Microsoft Indexing Service.

OLE DB Data Link Connections

Data Link File - UDL

strConnection = _T("File Name=c:\\DataBasePath\\DatabaseName.udl;");

如何提高数据库查询速度（收藏+体会）

肉丝 — Tue, 14 Jun 2005 08:13:00 GMT

1、用程序中，
保证在实现功能的基础上，尽量减少对数据库的访问次数；
通过搜索参数，尽量减少对表的访问行数,最小化结果集，从而减轻网络负担；
能够分开的操作尽量分开处理，提高每次的响应速度；
在数据窗口使用SQL时，尽量把使用的索引放在选择的首列；
算法的结构尽量简单；
在查询时，不要过多地使用通配符如SELECT * FROM T1语句，要用到几列就选择几列如：SELECT COL1,COL2 FROM T1；
在可能的情况下尽量限制尽量结果集行数如：SELECT TOP 300 COL1,COL2,COL3 FROM T1,因为某些情况下用户是不需要那么多的数据的。
不要在应用中使用数据库游标，游标是非常有用的工具，但比使用常规的、面向集的SQL语句需要更大的开销；
按照特定顺序提取数据的查找。

2、避免使用不兼容的数据类型。例如float和int、char和varchar、binary和varbinary是不兼容的。
数据类型的不兼容可能使优化器无法执行一些本来可以进行的优化操作。例如:
SELECT name FROM employee WHERE salary ＞ 60000
在这条语句中,如salary字段是money型的,则优化器很难对其进行优化,因为60000是个整型数。
我们应当在编程时将整型转化成为钱币型,而不要等到运行时转化。

3、尽量避免在WHERE子句中对字段进行函数或表达式操作，这将导致引擎放弃使用索引而进行全表扫描。如：
SELECT * FROM T1 WHERE F1/2=100
应改为:
SELECT * FROM T1 WHERE F1=100*2

SELECT * FROM RECORD WHERE SUBSTRING(CARD_NO,1,4)=’5378’
应改为:
SELECT * FROM RECORD WHERE CARD_NO LIKE ‘5378%’

SELECT member_number, first_name, last_name FROM members
WHERE DATEDIFF(yy,datofbirth,GETDATE()) > 21
应改为:
SELECT member_number, first_name, last_name FROM members
WHERE dateofbirth < DATEADD(yy,-21,GETDATE())
即：任何对列的操作都将导致表扫描，它包括数据库函数、计算表达式等等，查询时要尽可能将操作移至等号右边。

4、避免使用!=或＜＞、IS NULL或IS NOT NULL、IN ，NOT IN等这样的操作符,
因为这会使系统无法使用索引,而只能直接搜索表中的数据。例如:
SELECT id FROM employee WHERE id != 'B%'
优化器将无法通过索引来确定将要命中的行数,因此需要搜索该表的所有行。

5、尽量使用数字型字段，一部分开发人员和数据库管理人员喜欢把包含数值信息的字段设计为字符型，
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回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。

6、合理使用EXISTS,NOT EXISTS子句。如下所示：
1.SELECT SUM(T1.C1)FROM T1 WHERE(
(SELECT COUNT(*)FROM T2 WHERE T2.C2=T1.C2>0)
2.SELECT SUM(T1.C1) FROM T1WHERE EXISTS(
SELECT * FROM T2 WHERE T2.C2=T1.C2)
两者产生相同的结果，但是后者的效率显然要高于前者。因为后者不会产生大量锁定的表扫描或是索引扫描。
如果你想校验表里是否存在某条纪录，不要用count(*)那样效率很低，而且浪费服务器资源。可以用EXISTS代替。如：
IF (SELECT COUNT(*) FROM table_name WHERE column_name = 'xxx')
可以写成：
IF EXISTS (SELECT * FROM table_name WHERE column_name = 'xxx')

经常需要写一个T_SQL语句比较一个父结果集和子结果集，从而找到是否存在在父结果集中有而在子结果集中没有的记录，如：
1.SELECT a.hdr_key FROM hdr_tbl a---- tbl a 表示tbl用别名a代替
WHERE NOT EXISTS (SELECT * FROM dtl_tbl b WHERE a.hdr_key = b.hdr_key)

2.SELECT a.hdr_key FROM hdr_tbl a
LEFT JOIN dtl_tbl b ON a.hdr_key = b.hdr_key WHERE b.hdr_key IS NULL

3.SELECT hdr_key FROM hdr_tbl
WHERE hdr_key NOT IN (SELECT hdr_key FROM dtl_tbl)
三种写法都可以得到同样正确的结果，但是效率依次降低。

7、尽量避免在索引过的字符数据中，使用非打头字母搜索。这也使得引擎无法利用索引。
见如下例子：
SELECT * FROM T1 WHERE NAME LIKE ‘%L%’
SELECT * FROM T1 WHERE SUBSTING(NAME,2,1)=’L’
SELECT * FROM T1 WHERE NAME LIKE ‘L%’
即使NAME字段建有索引，前两个查询依然无法利用索引完成加快操作，引擎不得不对全表所有数据逐条操作来完成任务。
而第三个查询能够使用索引来加快操作。

8、充分利用连接条件，在某种情况下，两个表之间可能不只一个的连接条件，
这时在 WHERE 子句中将连接条件完整的写上，有可能大大提高查询速度。例：
SELECT SUM(A.AMOUNT) FROM ACCOUNT A,CARD B WHERE A.CARD_NO = B.CARD_NO
SELECT SUM(A.AMOUNT) FROM ACCOUNT A,CARD B WHERE A.CARD_NO = B.CARD_NO AND A.ACCOUNT_NO=B.ACCOUNT_NO
第二句将比第一句执行快得多。

9、消除对大型表行数据的顺序存取，尽管在所有的检查列上都有索引，但某些形式的WHERE子句强迫优化器使用顺序存取。如：
SELECT * FROM orders WHERE (customer_num=104 AND order_num>1001) OR order_num=1008
解决办法可以使用并集来避免顺序存取：
SELECT ＊ FROM orders WHERE customer_num=104 AND order_num>1001
UNION
SELECT ＊ FROM orders WHERE order_num=1008
这样就能利用索引路径处理查询。

10、避免困难的正规表达式。LIKE关键字支持通配符匹配，技术上叫正规表达式。但这种匹配特别耗费时间。例如：
SELECT * FROM customer WHERE zipcode LIKE “98_ _ _”
即使在zipcode字段上建立了索引，在这种情况下也还是采用顺序扫描的方式。如
果把语句改为SELECT ＊ FROM customer WHERE zipcode >“98000”，在执行查询
时就会利用索引来查询，显然会大大提高速度。

11、使用视图加速查询。把表的一个子集进行排序并创建视图，有时能加速查询。
它有助于避免多重排序操作，而且在其他方面还能简化优化器的工作。例如：
SELECT cust.name，rcvbles.balance，……other columns
FROM cust，rcvbles
WHERE cust.customer_id = rcvlbes.customer_id
AND rcvblls.balance>0
AND cust.postcode>“98000”
ORDER BY cust.name
如果这个查询要被执行多次而不止一次，可以把所有未付款的客户找出来放在一个
视图中，并按客户的名字进行排序：
CREATE VIEW DBO.V_CUST_RCVLBES
AS
SELECT cust.name，rcvbles.balance，……other columns
FROM cust，rcvbles
WHERE cust.customer_id = rcvlbes.customer_id
AND rcvblls.balance>0
ORDER BY cust.name

然后以下面的方式在视图中查询：
SELECT ＊ FROM V_CUST_RCVLBES
WHERE postcode>“98000”
视图中的行要比主表中的行少，而且物理顺序就是所要求的顺序，减少了磁盘I/O，所以查询工作量可以得到大幅减少。

12、能够用BETWEEN的就不要用IN
SELECT * FROM T1 WHERE ID IN (10,11,12,13,14)
改成：
SELECT * FROM T1 WHERE ID BETWEEN 10 AND 14
因为IN会使系统无法使用索引,而只能直接搜索表中的数据。

13、DISTINCT的就不用GROUP BY
SELECT OrderID FROM Details WHERE UnitPrice > 10 GROUP BY OrderID
可改为：
SELECT DISTINCT OrderID FROM Details WHERE UnitPrice > 10

14、部分利用索引
查询一：
SELECT employeeID, firstname, lastname
FROM names
WHERE dept = 'prod' or city = 'Orlando' or division = 'food'
查询二：
SELECT employeeID, firstname, lastname FROM names WHERE dept = 'prod'
UNION ALL
SELECT employeeID, firstname, lastname FROM names WHERE city = 'Orlando'
UNION ALL
SELECT employeeID, firstname, lastname FROM names WHERE division = 'food'
如果dept列建有索引则查询二可以部分利用索引，查询一则不能。

数据库

NOT IN 与NULL的比较

用ODBC方式读取EXCEL时，读取出来的内容无效（#错误)之类的解决方法。

ADO连接字符串

如何提高数据库查询速度 （收藏+体会）

如何提高数据库查询速度（收藏+体会）