日常使用SqlServer的笔记

表操作

• 查表大小

  sp_spaceused 表名

• 修改表名

  ALTER TABLE table RENAME TO NEW_TABLE_NAME;

• 修改列名
  不支持

  ALTER TABLE table RENAME COLUMN OLD_COLUMN_NAME TO NEW_COLUMN_NAME

  支持以下两种方式

  sp_rename 'change.Times','TIMES1','column'
  go
  -- 或者
  EXEC sp_rename 'T_WeiBoHot.citme','ctime','column'
  go
  -- change是表名，times是原来的列名，times1是新的列名

• 修改列的数据类型

  ALTER TABLE dbo.T_Intervene_HistoryDBRate ALTER COLUMN rate REAL;

• 插入/删除列

  ALTER TABLE table ADD COLUMN_NAME DATATYPE;
  go
  ALTER TABLE table DROP COLUMN COLUMN_NAME;

• 分区
  建分区

  CREATE PARTITION FUNCTION [fq_function](tinyint) AS RANGE LEFT FOR VALUES (0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07)
  GO
  CREATE PARTITION SCHEME [fq_Scheme] AS PARTITION [fq_function] TO ([PRIMARY], [PRIMARY], [PRIMARY], [PRIMARY], [PRIMARY], [PRIMARY], [PRIMARY], [PRIMARY], [PRIMARY])
  GO

  查询某分区数据

  SELECT * FROM dbo.T_monnewssource_FQTest WITH(NOLOCK) WHERE $PARTITION.fq31_function(fq_date) = 30;

  查当天数据所在分区

  SELECT $PARTITION.fq31_function(datediff(day,'2015-07-01',getdate())%31);

  分区依据

  fq_date TINYINT DEFAULT (datediff(day,'2015-01-01',getdate())%(8))
  -- 分区依据字段值从0~7，分区从1~8
  -- 31个分区时，只保留7个分区
  DECLARE @i TINYINT = 1,@j TINYINT = 25,@str VARCHAR(100) = '';
  WHILE @i < 25
  BEGIN
  SELECT @i = @i + 1;
  SELECT @str = @str + ',' + LTRIM(@j);
  IF @j = 31
  BEGIN
  SELECT @j = 0
  END
  SELECT @j = @j + 1;
  END
  
  SELECT STUFF(@str,1,1,'') AS strs;

过程，作业和触发器等操作

• 触发器

  禁用触发器：

  ALTER TABLE trig_example DISABLE TRIGGER trig1

  恢复触发器：

  ALTER TABLE trig_example ENABLE TRIGGER trig1

  禁用某个表上的所有触发器

  ALTER TABLE 你的表 DISABLE TRIGGER all

  启用某个表上的所有触发器

  ALTER TABLE 你的表 enable TRIGGER all

  禁用所有表上的所有触发器

  exec sp_msforeachtable 'ALTER TABLE ? DISABLE TRIGGER all'

  启用所有表上的所有触发器

  exec sp_msforeachtable 'ALTER TABLE ? enable TRIGGER all'

• 查询作业

  ``sql select a.name,b.command from msdb.dbo.sysjobs a inner join msdb.dbo.sysjobsteps b on a.job_id = b.job_id where b.command like '%P_GZ_HeavyPaste_UpOrder%';

• 查询过程

  select a.name,b.text
  from sys.objects a
  inner join sys.syscomments b
  on a.object_id = b.id
  where a.type = 'P'
  and b.text like '%T_SEM_PostSnap%';

• 查询触发器

  SELECT * FROM sys.objects WHERE type = 'tr';

• 查询视图

  select * from sysobjects where xtype='V';

• 查看进程

  SELECT PRO.spid
      ,DB_NAME(PRO.dbid) AS DBName
      ,PRO.program_name
      ,waitresource
      ,PRO.loginame AS LoginName
      ,DB.name AS DatabaseName
      ,PRO.[status] as ProcessStatus
      ,PRO.cmd AS Command
      ,PRO.last_batch AS LastBatch
      ,PRO.cpu AS Cpu
      ,PRO.physical_io AS PhysicalIo
      ,PRO.hostname
      ,SES.row_count AS [RowCount]
      ,STM.[text] AS SQLStatement
  FROM sys.sysprocesses AS PRO
      INNER JOIN sys.databases AS DB
          ON PRO.dbid = DB.database_id
      INNER JOIN sys.dm_exec_sessions AS SES
          ON PRO.spid = SES.session_id
      CROSS APPLY sys.dm_exec_sql_text(PRO.sql_handle) AS STM
  WHERE PRO.spid >= 50  -- Exclude system processes
  ORDER BY PRO.physical_io DESC ,PRO.cpu DESC
  go
  
  SELECT *
  FROM sys.sysprocesses AS pro
  CROSS APPLY sys.dm_exec_sql_text(sql_handle) AS sql
  WHERE pro.spid = 135

• 循环执行过程

  DECLARE @i INT =1
  WHILE @i=1
  BEGIN
  EXEC @i=P_wenda_ExportQuestionDaily; --过程return 1
  END

• 查看过程执行情况

  SELECT ISNULL(DBS.name, '') AS DatabaseName
      ,OBJECT_NAME(EPS.object_id, EPS.database_id) AS ObjectName
      ,EPS.cached_time AS CachedTime
      ,EPS.last_elapsed_time AS LastElapsedTime
      ,EPS.execution_count AS ExecutionCount
      ,EPS.total_worker_time / EPS.execution_count AS AvgWorkerTime
      ,EPS.total_elapsed_time / EPS.execution_count AS AvgElapsedTime
      ,(EPS.total_logical_reads + EPS.total_logical_writes)
      / EPS.execution_count AS AvgLogicalIO
  FROM sys.dm_exec_procedure_stats AS EPS
  LEFT JOIN sys.databases AS DBS
      ON EPS.database_id = DBS.database_id
  ORDER BY AvgWorkerTime DESC;

• 查看开销较大过程

  SELECT  ss.SUM_execution_count ,
      t.text ,
      ss.SUM_total_elapsed_time ,
      ss.sum_total_worker_time ,
      ss.sum_total_logical_reads ,
      ss.sum_total_logical_writes
  FROM    ( SELECT    S.plan_handle ,
                  SUM(s.execution_count) SUM_Execution_count ,
                  SUM(s.total_elapsed_time) SUM_total_elapsed_time ,
                  SUM(s.total_worker_time) SUM_total_worker_time ,
                  SUM(s.total_logical_reads) SUM_total_logical_reads ,
                  SUM(s.total_logical_writes) SUM_total_logical_writes
          FROM      sys.dm_exec_query_stats s
          GROUP BY  S.plan_handle
      ) AS ss
      CROSS APPLY sys.dm_exec_sql_text(ss.plan_handle) t
  ORDER BY sum_total_logical_reads DESC

连接

• full join

  SELECT (CASE WHEN A.customerId IS NULL THEN B.customerId ELSE A.customerId END) AS customerId,
  A.itemMoney AS currentNum, B.itemMoney AS frontNum
  INTO #W
  FROM #CA AS A
  FULL JOIN (SELECT customerId, itemMoney FROM #CA WHERE customerType = 2) AS B
  ON A.customerId = B.customerId
  WHERE A.customerType = 1
  go
  
  SELECT (CASE WHEN W.customerId IS NULL THEN C.customerId ELSE W.customerId END) AS customerId,
  W.currentNum, W.frontNum, C.itemMoney AS lastYearNum
  INTO #WW
  FROM #W AS W
  FULL JOIN (SELECT customerId, itemMoney FROM #CA WHERE customerType = 3) AS C
  ON W.customerId = C.customerId
  go

• cross apply

  SELECT F.usergroupid, F.row, S.Element
  FROM dbo.T_Cabin_FeatureWords AS F
  CROSS APPLY dbo.ufn_split(row, '-') AS S
  WHERE usergroupid = 21

• merge

  MERGE dbo.T_I8_CommentGroup AS A
  USING (SELECT @e_id AS id) AS B
  ON A.ug_id = B.id
  WHEN MATCHED THEN
  UPDATE SET ug_useflag = @ug_useflag, os_depict = @os_depict
  WHEN NOT MATCHED THEN
  INSERT (ug_id, ug_useflag, os_depict, lasttime)
  VALUES (@e_id, @ug_useflag, @os_depict, CONVERT(VARCHAR(10), GETDATE() - 7, 121))
  go
  
  MERGE dbo.T_I8_CommentUserGroup AS A
  USING
  (
  SELECT @e_id AS id, B.col
  FROM dbo.f_split(@ug_ids,',') AS B
  WHERE B.col > ''
  )AS C
  ON A.e_id = C.id AND A.ug_id = CONVERT(INT, C.col)
  WHEN NOT MATCHED THEN
  INSERT (e_id, ug_id)
  VALUES (C.id, C.col)
  WHEN NOT MATCHED BY SOURCE AND A.e_id = @e_id
  THEN DELETE
  go

索引

• 创建唯一索引并忽略重复值（已有的）

  CREATE UNIQUE INDEX idx_temp_id ON dbo.temp_lxc (id) WITH(IGNORE_DUP_KEY = ON);

• 删除索引语句

  DROP INDEX sy ON salary;

• 指定使用某索引

SELECT 字段名表  FROM 表名表 WITH (INDEX(索引名)) WHERE 查询条件;

• 建全文索引
  开启全文索引

  execute sp_fulltext_database 'enable';

  定义全文索引目录（fulltext_bbsBoard_PostDay）

  EXEC [sys].[sp_fulltext_catalog]
  @ftcat = 'fulltext_bbsBoard_PostDay', -- sysname
  @action = 'create', -- varchar(20)
  @path = N'D:\fulltext' -- nvarchar(101)

  为表[T_bbsBoard_PostDay]创建全文索引 可索引列为 id

  EXEC [sys].[sp_fulltext_table] @tabname = N'T_bbsBoard_PostDay', -- nvarchar(517)
  @action = 'create', -- varchar(50)
  @ftcat = fulltext_bbsBoard_PostDay, -- sysname
  @keyname = PK__T_bbsBoa__521F78666482D9EB -- sysname

  设置全文索引列名

  EXEC [sys].[sp_fulltext_column] @tabname = N'T_bbsBoard_PostDay', -- nvarchar(517)
  @colname = title, -- sysname
  @action = 'add', -- varchar(20)
  @language = 2052 -- int 2057 is the LCID for 英语 1033 :美语 2052：中文

  激活索引

  EXEC [sys].[sp_fulltext_table] @tabname = N'T_bbsBoard_PostDay', -- nvarchar(517)
  @action = 'activate'

  填充索引，首次创建完全文索引之后先完全填充索引，把数据全部放入全文索引里

  ALTER FULLTEXT INDEX ON dbo.T_bbsBoard_PostDay SET CHANGE_TRACKING OFF
  GO
  ALTER FULLTEXT INDEX ON dbo.T_bbsBoard_PostDay START FULL POPULATION
  GO

  修改填充方式为自动跟踪更改，等数据有变化的时候只需要把变化的部分填充进去全文索引里

  ALTER FULLTEXT INDEX ON dbo.T_bbsBoard_PostDay SET CHANGE_TRACKING AUTO
  GO

• 使用全文索引

WHERE CONTAINS(Title,'华为 and/or 联想')

约束

• 建唯一约束

  ALTER TABLE T_WenDa_QuestionZFM ADD CONSTRAINT un_QuestionZFM_autoid UNIQUE(autoid);
  go
  ALTER TABLE dbo.T_MS_TopicBlackList ADD CONSTRAINT pk_TopicBlackList_topicid PRIMARY KEY (topic_id);

• 删约束

  ALTER TABLE dbo.T_SEM_TriLo DROP CONSTRAINT DF__T_SEM_Tri__ji_fl__4262CC11;

• 为已有列添加默认值

  ALTER TABLE dbo.T_SEM_ShowJsonShotSnap ADD CONSTRAINT ShowJsonShotSnap_defualt_rgNum DEFAULT(0) FOR rgNum;

时间日期处理

• 显示年月日部分：

  convert(nvarchar(10), date1, 21)

• 取得该月第一天：

  DATEADD(DD, -DAY(GETDATE()) + 1, GETDATE())

• 取得该月最后一天：

  DATEADD(DD, -DAY(GETDATE()), DATEADD(MM, 1, GETDATE()))

算法

• 返回插入到标识列中的值(SCOPE_IDENTITY/IDENT_CURRENT/@@IDENTITY)
  返回插入数据条数使用 @@ROWCOUNT

  CREATE TABLE #A(id INT);
  INSERT INTO dbo.tongxuelu
          ( t_name, t_sex, t_age )
  OUTPUT INSERTED.t_id INTO #A
  SELECT 'abc','女',0;

• 遍历

  ;
  WITH SUB(id, fid, word, paths,layer,word_id) AS (
  SELECT C.cls_id,C.cls_fid,C.cls_path_word,C.cls_path_wordstr,C.cls_layer,C.cls_path_id
  FROM QADATA.dbo.T_QA_Class AS C WITH(NOLOCK) WHERE C.cls_id = 868163
  UNION ALL
  SELECT C.cls_id,C.cls_fid,C.cls_path_word,C.cls_path_wordstr,C.cls_layer,C.cls_path_id
  FROM SUB AS S
  INNER JOIN QADATA.dbo.T_QA_Class AS C WITH(NOLOCK) ON S.id = C.cls_fid
  )
  SELECT * FROM SUB ORDER BY layer ASC;

• 随机函数 NEWID()

• 组内排序

  ROW_NUMBER() OVER(PARTITION BY A.Url ORDER BY A.Utime DESC) AS n

• 哈希排重（MD5）

  SELECT CONVERT(UNIQUEIDENTIFIER,HASHBYTES('MD5','LXC'));
  SELECT @titlehash = CONVERT(UNIQUEIDENTIFIER,HASHBYTES('md5', SUBSTRING( ISNULL(@title, ''),1,255)));
  contenthash = CONVERT(UNIQUEIDENTIFIER , HASHBYTES('md5',LEFT(RTRIM(LTRIM(@content)),4000))

• 拼接字符串

  SELECT t_type + ',' FROM #toolList FOR XML PATH('');
  go
  
  CREATE TABLE #words(id INT,word NVARCHAR(8));
  INSERT INTO #words( id,word )
  SELECT 1,'过去'
  UNION
  SELECT 1,'现在'
  UNION
  SELECT 1,'将来'
  UNION
  SELECT 2,'过去2'
  UNION
  SELECT 3,'现在3'
  UNION
  SELECT 2,'将来2'
  
  SELECT * FROM #words;
  
  DECLARE @a NVARCHAR(max);
  
  SELECT @a = STUFF((SELECT '|' + word FROM #words FOR XML PATH('')),1,1,'');
  SELECT @a;
  go
  
  SELECT B.id,STUFF((SELECT '|' + word FROM #words AS A WHERE A.id = B.id FOR XML PATH('')),1,1,'')
  FROM #words AS B
  GROUP BY B.id
  
  DROP TABLE #words;
  
  CREATE TABLE #temp(id INT);
  
  INSERT INTO #temp
          ( id )
  SELECT 1
  UNION
  SELECT 2
  UNION
  SELECT 3;
  
  SELECT STUFF((SELECT ',' + LTRIM(id) FROM #temp FOR XML PATH('')),1,1,'');
  DROP TABLE #temp;

• 平均分配（个数/时间）

  CREATE TABLE #temp(id INT,iname VARCHAR(10));
  INSERT INTO #temp( id, iname )
  SELECT 1,'a'
  UNION
  SELECT 2,'b'
  UNION
  SELECT 3,'c'
  
  SELECT * FROM #temp;
  
  SELECT * FROM #temp AS t,dbo.Nums AS n WHERE n.n <= CEILING(10*1.0/3);
  
  DECLARE @endTime DATETIME = GETDATE() + 1;
  SELECT *,DATEADD(MINUTE,(id-1) * DATEDIFF(MINUTE,GETDATE(),@endTime) / 4,GETDATE())
  FROM #temp

• 计算某词在串中出现个数

  DECLARE @str VARCHAR(100)= 'a,b,c,d'
  SELECT @str
  SELECT LEN(@str) - LEN(REPLACE(@str, 'a', '')) 

• 构造日期表

  SELECT CAST(GETDATE()-number AS DATE) AS time
  FROM master..spt_values WHERE type='P'
  AND number<30

• 构造递增表

  SELECT number FROM master..spt_values WHERE type = 'P' AND number < 24

• 判断字符串是否是纯数字或字母

  -- 纯数字：
  ISNUMERIC(title) = 1
  -- 纯数字：
  PATINDEX('%[^0-9]%','12324') = 0
  -- 纯英文：
  PATINDEX('%[^a-z]%','abc') = 0
  --- 不含汉字和字母：
  PATINDEX('%[吖-咗a-zA-Z]%',Title) = 0
  -- 不含汉字：
  PATINDEX('%[吖-咗]%',title) = 0

• 正则表达式

  获取字符串中的英文字符

  declare @str varchar(100)
  set @str='xyz123z5'
  select dbo.regexReplace(@str,'[^a-zA-Z]','',1,1)

  获取字符串中的英文字符长度

  declare @str varchar(100)
  set @str='xyz123z5'
  select len(dbo.regexReplace(@str,'[^a-zA-Z]','',1,1))

  语句块例子：

  create function dbo.regexReplace
  (
  @source varchar(5000), --原字符串
  @regexp varchar(1000), --正则表达式
  @replace varchar(1000), --替换值
  @globalReplace bit = 0, --是否是全局替换
  @ignoreCase bit = 0 --是否忽略大小写
  )
  returnS varchar(1000) AS
  begin
  declare @hr integer
  declare @objRegExp integer
  declare @result varchar(5000)
  
  exec @hr = sp_OACreate 'VBScript.RegExp', @objRegExp OUTPUT
  IF @hr <> 0 begin
  exec @hr = sp_OADestroy @objRegExp
  return null
  end
  exec @hr = sp_OASetProperty @objRegExp, 'Pattern', @regexp
  IF @hr <> 0 begin
  exec @hr = sp_OADestroy @objRegExp
  return null
  end
  exec @hr = sp_OASetProperty @objRegExp, 'Global', @globalReplace
  IF @hr <> 0 begin
  exec @hr = sp_OADestroy @objRegExp
  return null
  end
  exec @hr = sp_OASetProperty @objRegExp, 'IgnoreCase', @ignoreCase
  IF @hr <> 0 begin
  exec @hr = sp_OADestroy @objRegExp
  return null
  end 
  exec @hr = sp_OAMethod @objRegExp, 'Replace', @result OUTPUT, @source, @replace
  IF @hr <> 0 begin
  exec @hr = sp_OADestroy @objRegExp
  return null
  end
  exec @hr = sp_OADestroy @objRegExp
  IF @hr <> 0 begin
  return null
  end
  
  return @result
  end
  GO

其他

• 创建数据库

  create database stuDB 
  on  primary  -- 默认就属于primary文件组,可省略
  (
  /*--数据文件的具体描述--*/
      name='stuDB_data',  -- 主数据文件的逻辑名称
      filename='D:\stuDB_data.mdf', -- 主数据文件的物理名称
      size=5mb, --主数据文件的初始大小
      maxsize=100mb, -- 主数据文件增长的最大值，不限制的话可用“unlimited”
      filegrowth=15%--主数据文件的增长率，也可指定大小
  )
  log on
  (
  /*--日志文件的具体描述,各参数含义同上--*/
      name='stuDB_log',
      filename='D:\stuDB_log.ldf',
      size=2mb,
      filegrowth=1mb --也可用增长率
  )

  一个数据库中，只可以有一个 主数据文件（扩展名： .mdf），多个次数据文件(扩展名： .ndf)，多个日志文件(扩展名： .Ldf)；

  多文件及文件组示例：

  CREATE DATABASE Sales
  ON PRIMARY
  ( NAME = SPri1_dat,
      FILENAME = 'D:\SalesData\SPri1dat.mdf',
      SIZE = 10,
      MAXSIZE = 50,
      FILEGROWTH = 15% ),
  ( NAME = SPri2_dat,
      FILENAME = 'D:\SalesData\SPri2dt.ndf',
      SIZE = 10,
      MAXSIZE = 50,
      FILEGROWTH = 15% ),
  FILEGROUP SalesGroup1
  ( NAME = SGrp1Fi1_dat,
      FILENAME = 'D:\SalesData\SG1Fi1dt.ndf',
      SIZE = 10,
      MAXSIZE = 50,
      FILEGROWTH = 5 ),
  ( NAME = SGrp1Fi2_dat,
      FILENAME = 'D:\SalesData\SG1Fi2dt.ndf',
      SIZE = 10,
      MAXSIZE = 50,
      FILEGROWTH = 5 ),
  FILEGROUP SalesGroup2
  ( NAME = SGrp2Fi1_dat,
      FILENAME = 'D:\SalesData\SG2Fi1dt.ndf',
      SIZE = 10,
      MAXSIZE = 50,
      FILEGROWTH = 5 ),
  ( NAME = SGrp2Fi2_dat,
      FILENAME = 'D:\SalesData\SG2Fi2dt.ndf',
      SIZE = 10,
      MAXSIZE = 50,
      FILEGROWTH = 5 )
  LOG ON
  ( NAME = Sales_log,
      FILENAME = 'E:\SalesLog\salelog.ldf',
      SIZE = 5MB,
      MAXSIZE = 25MB,
      FILEGROWTH = 5MB ) ;
  GO

• 关闭/开启 Transact-SQL 语句或存储过程影响的行计数的消息

  SET NOCOUNT ON
  go
  SET NOCOUNT OFF
  go
  SET NOCOUNT ON

• 符号
  回车符：char(13)
  换行符：char(10)
  制表符： CHAR(9)

• 开启/关闭自增列
  显示值插入(修改会话中的IDENTITY_INSERT ),临时性 ,不彻底改表列性质

  SET IDENTITY_INSERT [ database_name . [ schema_name ] . ] table { ON | OFF }

• 查服务器名

  SELECT @@SERVERNAME;

• 页面无法执行过程时，环境设置

  SET ANSI_WARNINGS ON;
  SET ANSI_PADDING ON;
  SET CONCAT_NULL_YIELDS_NULL ON;

• 锁
  查死锁

  select   request_session_id   spid,OBJECT_NAME(resource_associated_entity_id) tableName  
  from   sys.dm_tran_locks where resource_type='OBJECT'
  AND OBJECT_NAME(resource_associated_entity_id) = 'MonNews_floors_ContentText';
  
  select   request_session_id   spid,OBJECT_NAME(resource_associated_entity_id) tableName  
  from   sys.dm_tran_locks where resource_type='OBJECT';
  
  select top 100* from master..sysprocesses a cross apply sys.dm_exec_sql_text(a.sql_handle) b
  where a.spid>50 AND b.text LIKE '%T_bbsBoard_PostDay%' order by hostname;
  
  select   request_session_id spid,OBJECT_NAME(resource_associated_entity_id) tableName   
  from sys.dm_tran_locks  
  where resource_type='OBJECT;

  解锁

  declare @spid  int
  Set @spid  = 57 --锁表进程
  declare @sql varchar(1000)
  set @sql='kill '+cast(@spid  as varchar)
  exec(@sql)

  检索死锁进程

  select spid, blocked, loginame, last_batch, status, cmd, hostname, program_name
  from sysprocesses
  where spid in
  (select blocked from sysprocesses where blocked <> 0 ) or (blocked <>0)

• 表A数据量很大,现需要保留表A的少量数据
  删除A耗时长而且会增加数据库负载，可以考虑更换表名的方法达到目的
  假设表A有自增id

  1).新建与表 A 结构相同的表 B；
  2).停止表 A 的作业修改过程 让表A数据不再变化；
  3).将 表 A 要保留数据 指定相同自增id 插入 表 B；

  SET IDENTITY_INSERT B ON
  INSERT B
  SELECT ... FROM A WHERE ...
  SET IDENTITY_INSERT B OFF

  4).指定 表 B 自增id；
  查询 表 A 自增id当前值DBCC CHECKIDENT (A)；
  指定表 B 为 与 表 A 一样的新值DBCC CHECKIDENT (B, RESEED, value)；
  5).重命名的方式到达 备份删除目的；

  EXEC sp_rename A,A1
  EXEC sp_rename B,A

• 查询相邻两条记录间的数据增加或减少值
  通过查询原来的数据库，建立两个临时表。临时表中添加一个字段作为顺序标签。
  第一个临时表中顺序标签是0开始，第二个临时表从1开始。然后两个表直接通过顺序标签这一列进行内联，这样不就能直接计算增量值了啊。在查询的时候还能直接把其他相关的时间等信息都带出来。

  注意，这样是让两个表格对应错位，以便求增量，所以两个临时表中都有一行在内联中丢失掉。