18

我有一个在“位置 A”中更改的查找数据表,我需要更新“位置 B”中相同的命名表。我在想,如果我有办法将表格(80k 行)逐行读取为格式正确的字符串,并使用合并语句创建输出文件(将表格“转换”为 CTE),那么我可以有一些光滑的东西。

对这个想法的所有想法(或那些已经做过类似事情的人的指点)都表示赞赏。谢谢。我正在使用 SQL Server Express 2012。

4

3 回答 3

29

我使用这个帮助脚本来管理环境之间的小型查找表。80k 行无论如何都不算​​大,但也不是微不足道的。康拉德为移动大量数据制定了可靠的计划。

如果您确实选择使用生成的合并路线,希望这可以帮助您:

declare @TableName nvarchar(128) = 'dbo.YourTable';

--// do not modify
set nocount on; 

if  (   isnull(parsename(@TableName, 1), '') = '' or
        isnull(parsename(@TableName, 2), '') = ''
    )
begin
    raiserror('@TableName should be in form of owner.object (dbo.Yak). Cannot continue.', 16, 1, @TableName);
    return;
end
if object_id(@TableName) is null
begin
    raiserror('Table [%s] not found. Cannot continue.', 16, 1, @TableName);
    return;
end
if objectproperty(object_id(@TableName), N'TableHasPrimaryKey') = 0
begin
    raiserror('Table [%s] has no primary key defined. Cannot continue.', 16, 1, @TableName);
    return;
end

declare @output varchar(max)='', @return int, @sql varchar(max) = '', @list varchar(max) = '', @exec nvarchar(max), @Cols varchar(max)='',
        @sCols varchar(max)='', @aCols varchar(max)='', @pCols varchar(max)='', @uCols varchar(max)='', 
        @b char(1) = char(13), @t char(1) = char(9);

select  @exec = isnull(@exec + '+' + @b,'') + ''','' +' +
        case 
            when c.data_type in ('uniqueidentifier') 
                then @t + '''''+isnull('''''''' + convert(varchar(50),' + c.column_name + ') + '''''''',''null'')'
            when c.data_type in ('xml') 
                then @t + '''''+isnull('''''''' + convert(varchar(max),' + c.column_name + ') + '''''''',''null'')'
            when c.data_type in ('char', 'nchar', 'varchar', 'nvarchar', 'sysname') 
                then @t + '''''+isnull('''''''' + replace(' + c.column_name + ','''''''','''''''''''') + '''''''',''null'')'
            when c.data_type in ('datetime', 'date') 
                then @t + '''''+isnull('''''''' + convert(varchar,' + c.column_name + ',121)+'''''''',''null'') '
            when c.data_type in ('tinyint', 'int', 'float', 'numeric', 'decimal', 'smallmoney', 'money', 'bit', 'smallint', 'real', 'bigint') 
                then @t + '''''+isnull(convert(varchar,' + c.column_name + '),''null'')' 
            else ''' ** data type [' + c.data_type + '] not supported **'''
        end,
        @Cols = @Cols + ','+quotename(c.column_name, '['),
        @sCols = @sCols + ',s.'+quotename(c.column_name, '[')
from    information_schema.columns c
where   c.table_name = parsename(@TableName, 1) and
        c.table_schema = parsename(@TableName, 2)
order 
by      c.ordinal_position

-- stage primary key columns
declare @pks table (c varchar(100), o int);
insert into @pks (c,o)
    select  kcu.column_name, kcu.ordinal_position
    from    information_schema.table_constraints tc
    join    information_schema.key_column_usage kcu on
            tc.table_name = kcu.table_name and
            tc.constraint_name = kcu.constraint_name
    where   tc.table_name = parsename(@TableName, 1) and
            tc.table_schema = parsename(@TableName, 2) and
            tc.constraint_type = 'PRIMARY KEY';
   
-- build primary key columns (1=1 to ease support of composite PKs)
set @pCols = '1=1'           
select  @pCols = @pCols + ' and '+isnull('t.'+quotename(c, '[') + ' = ' + 's.'+quotename(c, '['), '')
from    @pks    
order
by      o;

-- build update columns, do not update identities or pks
select  @aCols = @aCols + ','+quotename(c.[name], '[') + ' = s.' + quotename(c.[name], '[')
from    sys.columns c
where   object_id = object_id(@TableName) and
        [name] not in (select c from @pks) and
        columnproperty(object_id(@TableName), c.[name], 'IsIdentity ') = 0;

-- script the data out as table value constructors
select  @exec = 'set nocount on; Select ' + @b + '''(' + ''' + ' + @b + stuff(@exec,1, 3, '') + '+'')''' + @b + 'from ' + @TableName,
        @Cols = stuff(@Cols,1, 1, ''),
        @sCols = stuff(@sCols,1, 1, ''),
        @aCols = stuff(@aCols,1, 1, '');
        
declare @tab table (val varchar(max));
declare @Values varchar(max);
insert into @tab
    exec(@exec);

if not exists(select 1 from @tab)
begin
    raiserror('Table %s is valid but empty. Populate it before running this helper.', 16, 1, @TableName);
    return
end

select @Values = stuff(cast((select ','+ @b + val from @tab for xml path('')) as xml).value('.', 'varchar(max)'),1,2,'');

-- build the merge statement
set @output +=  @b+'--'+@TableName+replicate('-', 98-len(@TableName))+@b+replicate('-', 100)+@b;
set @output +=  'set nocount on;'+@b
if objectproperty(object_id(@TableName), 'TableHasIdentity') = 1 
    set @output += 'set identity_insert ['+parsename(@TableName, 2)+'].[' + parsename(@TableName, 1) + '] on;'+@b;
set @output +=  ';with cte_data('+@Cols+')'+@b+'as (select * from (values'+@b+'--'+replicate('/', 98) + @b +@Values+ @b +'--'+replicate('\', 98)+ @b +')c('+@Cols+'))'+@b;
set @output +=  'merge' + @t + '['+parsename(@TableName, 2)+'].[' + parsename(@TableName, 1) + '] as t' + @b + 'using' + @t + 'cte_data as s'+@b;
set @output +=  'on' + replicate(@t, 2) + @pCols+@b;
set @output +=  'when matched then' + @b+@t + 'update set'+ @b+@t + @aCols+@b;
set @output +=  'when not matched by target then'+@b;
set @output +=  @t+'insert(' + @Cols + ')'+@b;
set @output +=  @t+'values(' + @sCols + ')'+@b;
set @output +=  'when not matched by source then' + @b+@t+ 'delete;'+@b;
if objectproperty(object_id(@TableName), 'TableHasIdentity') = 1 
    set @output += 'set identity_insert ['+parsename(@TableName, 2)+'].[' + parsename(@TableName, 1) + '] off;'

--output the statement as xml (to overcome mgmt studio limitations)
select s as [output] from (select @output)d(s) for xml path('');
return;
于 2012-05-07T22:01:47.677 回答
18

此帮助程序脚本将表名作为参数并生成包含数据的 MERGE 语句。它将存储过程[master]作为系统对象安装在其中,允许从所有数据库(例如EXEC Northwind.dbo.sp_generate_merge 'Region')调用 proc

这种方法(最初是我在SQL Server Central上发布的)适用于长度超过 8,000 个字符的结果集,我们倾向于达到 100 行标记。

它还支持包含/排除特定列以及在源数据发生更改时有条件地更新匹配的行。

更新于2012-09-27更改了 proc 以使用 Nathan 改进的 Xml 输出方法来生成 MERGE 语句,从而解决了行和结果集级别的字符限制问题。

2013年 4 月 22 日更新,修复了 Dave 报告的 xml 字符问题

更新于2013-04-22这个脚本现在在 GitHub 上:https ://github.com/readyroll/generate-sql-merge

SET NOCOUNT ON
GO

PRINT 'Using Master database'
USE master
GO

PRINT 'Checking for the existence of this procedure'
IF (SELECT OBJECT_ID('sp_generate_merge','P')) IS NOT NULL --means, the procedure already exists
 BEGIN
 PRINT 'Procedure already exists. So, dropping it'
 DROP PROC sp_generate_merge
 END
GO

--Turn system object marking on

CREATE PROC sp_generate_merge
(
 @table_name varchar(776), -- The table/view for which the MERGE statement will be generated using the existing data
 @target_table varchar(776) = NULL, -- Use this parameter to specify a different table name into which the data will be inserted/updated/deleted
 @from varchar(800) = NULL, -- Use this parameter to filter the rows based on a filter condition (using WHERE)
 @include_timestamp bit = 0, -- Specify 1 for this parameter, if you want to include the TIMESTAMP/ROWVERSION column's data in the MERGE statement
 @debug_mode bit = 0, -- If @debug_mode is set to 1, the SQL statements constructed by this procedure will be printed for later examination
 @schema varchar(64) = NULL, -- Use this parameter if you are not the owner of the table
 @ommit_images bit = 0, -- Use this parameter to generate MERGE statement by omitting the 'image' columns
 @ommit_identity bit = 0, -- Use this parameter to ommit the identity columns
 @top int = NULL, -- Use this parameter to generate a MERGE statement only for the TOP n rows
 @cols_to_include varchar(8000) = NULL, -- List of columns to be included in the MERGE statement
 @cols_to_exclude varchar(8000) = NULL, -- List of columns to be excluded from the MERGE statement
 @update_only_if_changed bit = 1, -- When 1, only performs an UPDATE operation if an included column in a matched row has changed.
 @disable_constraints bit = 0, -- When 1, disables foreign key constraints and enables them after the MERGE statement
 @ommit_computed_cols bit = 0, -- When 1, computed columns will not be included in the MERGE statement
 @include_use_db bit = 1, -- When 1, includes a USE [DatabaseName] statement at the beginning of the generated batch
 @results_to_text bit = 0 -- When 1, outputs results to grid/messages window. When 0, outputs MERGE statement in an XML fragment.
)
AS
BEGIN

/***********************************************************************************************************
Procedure: sp_generate_merge (Version 0.93)
 (Adapted by Daniel Nolan for SQL Server 2008/2012)

Adapted from: sp_generate_inserts (Build 22) 
 (Copyright © 2002 Narayana Vyas Kondreddi. All rights reserved.)

Purpose: To generate a MERGE statement from existing data, which will INSERT/UPDATE/DELETE data based
 on matching primary key values in the source/target table.

 The generated statements can be executed to replicate the data in some other location.

Written by: Narayana Vyas Kondreddi
 http://vyaskn.tripod.com

 Daniel Nolan
 http://danere.com
 @dan3r3

Example 1: To generate a MERGE statement for table 'titles':

 EXEC sp_generate_merge 'titles'

Example 2: To generate a MERGE statement for 'titlesCopy' table from 'titles' table:

 EXEC sp_generate_merge 'titles', 'titlesCopy'

Example 3: To generate a MERGE statement for table 'titles' that will unconditionally UPDATE matching rows 
 (ie. not perform a "has data changed?" check prior to going ahead with an UPDATE):

 EXEC sp_generate_merge 'titles', @update_only_if_changed = 0

Example 4: To generate a MERGE statement for 'titles' table for only those titles 
 which contain the word 'Computer' in them:
 NOTE: Do not complicate the FROM or WHERE clause here. It's assumed that you are good with T-SQL if you are using this parameter

 EXEC sp_generate_merge 'titles', @from = "from titles where title like '%Computer%'"

Example 5: To specify that you want to include TIMESTAMP column's data as well in the MERGE statement:
 (By default TIMESTAMP column's data is not scripted)

 EXEC sp_generate_merge 'titles', @include_timestamp = 1

Example 6: To print the debug information:

 EXEC sp_generate_merge 'titles', @debug_mode = 1

Example 7: If the table is in a different schema to the default, use @schema parameter to specify the schema name
 To use this option, you must have SELECT permissions on that table

 EXEC sp_generate_merge 'Nickstable', @schema = 'Nick'

Example 8: To generate a MERGE statement for the rest of the columns excluding images

 EXEC sp_generate_merge 'imgtable', @ommit_images = 1

Example 9: To generate a MERGE statement excluding (omitting) IDENTITY columns:
 (By default IDENTITY columns are included in the MERGE statement)

 EXEC sp_generate_merge 'mytable', @ommit_identity = 1

Example 10: To generate a MERGE statement for the TOP 10 rows in the table:

 EXEC sp_generate_merge 'mytable', @top = 10

Example 11: To generate a MERGE statement with only those columns you want:

 EXEC sp_generate_merge 'titles', @cols_to_include = "'title','title_id','au_id'"

Example 12: To generate a MERGE statement by omitting certain columns:

 EXEC sp_generate_merge 'titles', @cols_to_exclude = "'title','title_id','au_id'"

Example 13: To avoid checking the foreign key constraints while loading data with a MERGE statement:

 EXEC sp_generate_merge 'titles', @disable_constraints = 1

Example 14: To exclude computed columns from the MERGE statement:

 EXEC sp_generate_merge 'MyTable', @ommit_computed_cols = 1

***********************************************************************************************************/

SET NOCOUNT ON


--Making sure user only uses either @cols_to_include or @cols_to_exclude
IF ((@cols_to_include IS NOT NULL) AND (@cols_to_exclude IS NOT NULL))
 BEGIN
 RAISERROR('Use either @cols_to_include or @cols_to_exclude. Do not use both the parameters at once',16,1)
 RETURN -1 --Failure. Reason: Both @cols_to_include and @cols_to_exclude parameters are specified
 END


--Making sure the @cols_to_include and @cols_to_exclude parameters are receiving values in proper format
IF ((@cols_to_include IS NOT NULL) AND (PATINDEX('''%''',@cols_to_include) = 0))
 BEGIN
 RAISERROR('Invalid use of @cols_to_include property',16,1)
 PRINT 'Specify column names surrounded by single quotes and separated by commas'
 PRINT 'Eg: EXEC sp_generate_merge titles, @cols_to_include = "''title_id'',''title''"'
 RETURN -1 --Failure. Reason: Invalid use of @cols_to_include property
 END

IF ((@cols_to_exclude IS NOT NULL) AND (PATINDEX('''%''',@cols_to_exclude) = 0))
 BEGIN
 RAISERROR('Invalid use of @cols_to_exclude property',16,1)
 PRINT 'Specify column names surrounded by single quotes and separated by commas'
 PRINT 'Eg: EXEC sp_generate_merge titles, @cols_to_exclude = "''title_id'',''title''"'
 RETURN -1 --Failure. Reason: Invalid use of @cols_to_exclude property
 END


--Checking to see if the database name is specified along wih the table name
--Your database context should be local to the table for which you want to generate a MERGE statement
--specifying the database name is not allowed
IF (PARSENAME(@table_name,3)) IS NOT NULL
 BEGIN
 RAISERROR('Do not specify the database name. Be in the required database and just specify the table name.',16,1)
 RETURN -1 --Failure. Reason: Database name is specified along with the table name, which is not allowed
 END


--Checking for the existence of 'user table' or 'view'
--This procedure is not written to work on system tables
--To script the data in system tables, just create a view on the system tables and script the view instead
IF @schema IS NULL
 BEGIN
 IF NOT EXISTS (SELECT 1 FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = @table_name AND (TABLE_TYPE = 'BASE TABLE' OR TABLE_TYPE = 'VIEW') AND TABLE_SCHEMA = SCHEMA_NAME())
 BEGIN
 RAISERROR('User table or view not found.',16,1)
 PRINT 'You may see this error if the specified table is not in your default schema (' + SCHEMA_NAME() + '). In that case use @schema parameter to specify the schema name.'
 PRINT 'Make sure you have SELECT permission on that table or view.'
 RETURN -1 --Failure. Reason: There is no user table or view with this name
 END
 END
ELSE
 BEGIN
 IF NOT EXISTS (SELECT 1 FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = @table_name AND (TABLE_TYPE = 'BASE TABLE' OR TABLE_TYPE = 'VIEW') AND TABLE_SCHEMA = @schema)
 BEGIN
 RAISERROR('User table or view not found.',16,1)
 PRINT 'Make sure you have SELECT permission on that table or view.'
 RETURN -1 --Failure. Reason: There is no user table or view with this name 
 END
 END


--Variable declarations
DECLARE @Column_ID int, 
 @Column_List varchar(8000), 
 @Column_List_For_Update varchar(8000), 
 @Column_List_For_Check varchar(8000), 
 @Column_Name varchar(128), 
 @Column_Name_Unquoted varchar(128), 
 @Data_Type varchar(128), 
 @Actual_Values nvarchar(max), --This is the string that will be finally executed to generate a MERGE statement
 @IDN varchar(128), --Will contain the IDENTITY column's name in the table
 @Target_Table_For_Output varchar(776),
 @Source_Table_Qualified varchar(776)



--Variable Initialization
SET @IDN = ''
SET @Column_ID = 0
SET @Column_Name = ''
SET @Column_Name_Unquoted = ''
SET @Column_List = ''
SET @Column_List_For_Update = ''
SET @Column_List_For_Check = ''
SET @Actual_Values = ''

--Variable Defaults
IF @schema IS NULL
 BEGIN
 SET @Target_Table_For_Output = QUOTENAME(COALESCE(@target_table, @table_name))
 END
ELSE
 BEGIN
 SET @Target_Table_For_Output = QUOTENAME(@schema) + '.' + QUOTENAME(COALESCE(@target_table, @table_name))
 END

SET @Source_Table_Qualified = QUOTENAME(COALESCE(@schema,SCHEMA_NAME())) + '.' + QUOTENAME(@table_name)

--To get the first column's ID
SELECT @Column_ID = MIN(ORDINAL_POSITION) 
FROM INFORMATION_SCHEMA.COLUMNS (NOLOCK) 
WHERE TABLE_NAME = @table_name
AND TABLE_SCHEMA = COALESCE(@schema, SCHEMA_NAME())


--Loop through all the columns of the table, to get the column names and their data types
WHILE @Column_ID IS NOT NULL
 BEGIN
 SELECT @Column_Name = QUOTENAME(COLUMN_NAME), 
 @Column_Name_Unquoted = COLUMN_NAME,
 @Data_Type = DATA_TYPE 
 FROM INFORMATION_SCHEMA.COLUMNS (NOLOCK) 
 WHERE ORDINAL_POSITION = @Column_ID
 AND TABLE_NAME = @table_name
 AND TABLE_SCHEMA = COALESCE(@schema, SCHEMA_NAME())

 IF @cols_to_include IS NOT NULL --Selecting only user specified columns
 BEGIN
 IF CHARINDEX( '''' + SUBSTRING(@Column_Name,2,LEN(@Column_Name)-2) + '''',@cols_to_include) = 0 
 BEGIN
 GOTO SKIP_LOOP
 END
 END

 IF @cols_to_exclude IS NOT NULL --Selecting only user specified columns
 BEGIN
 IF CHARINDEX( '''' + SUBSTRING(@Column_Name,2,LEN(@Column_Name)-2) + '''',@cols_to_exclude) <> 0 
 BEGIN
 GOTO SKIP_LOOP
 END
 END

 --Making sure to output SET IDENTITY_INSERT ON/OFF in case the table has an IDENTITY column
 IF (SELECT COLUMNPROPERTY( OBJECT_ID(@Source_Table_Qualified),SUBSTRING(@Column_Name,2,LEN(@Column_Name) - 2),'IsIdentity')) = 1 
 BEGIN
 IF @ommit_identity = 0 --Determing whether to include or exclude the IDENTITY column
 SET @IDN = @Column_Name
 ELSE
 GOTO SKIP_LOOP 
 END

 --Making sure whether to output computed columns or not
 IF @ommit_computed_cols = 1
 BEGIN
 IF (SELECT COLUMNPROPERTY( OBJECT_ID(@Source_Table_Qualified),SUBSTRING(@Column_Name,2,LEN(@Column_Name) - 2),'IsComputed')) = 1 
 BEGIN
 GOTO SKIP_LOOP 
 END
 END

 --Tables with columns of IMAGE data type are not supported for obvious reasons
 IF(@Data_Type in ('image'))
 BEGIN
 IF (@ommit_images = 0)
 BEGIN
 RAISERROR('Tables with image columns are not supported.',16,1)
 PRINT 'Use @ommit_images = 1 parameter to generate a MERGE for the rest of the columns.'
 RETURN -1 --Failure. Reason: There is a column with image data type
 END
 ELSE
 BEGIN
 GOTO SKIP_LOOP
 END
 END

 --Determining the data type of the column and depending on the data type, the VALUES part of
 --the MERGE statement is generated. Care is taken to handle columns with NULL values. Also
 --making sure, not to lose any data from flot, real, money, smallmomey, datetime columns
 SET @Actual_Values = @Actual_Values +
 CASE 
 WHEN @Data_Type IN ('char','nchar') 
 THEN 
 'COALESCE('''''''' + REPLACE(RTRIM(' + @Column_Name + '),'''''''','''''''''''')+'''''''',''NULL'')'
 WHEN @Data_Type IN ('varchar','nvarchar') 
 THEN 
 'COALESCE('''''''' + REPLACE(' + @Column_Name + ','''''''','''''''''''')+'''''''',''NULL'')'
 WHEN @Data_Type IN ('datetime','smalldatetime','datetime2') 
 THEN 
 'COALESCE('''''''' + RTRIM(CONVERT(char,' + @Column_Name + ',127))+'''''''',''NULL'')'
 WHEN @Data_Type IN ('uniqueidentifier') 
 THEN 
 'COALESCE('''''''' + REPLACE(CONVERT(char(255),RTRIM(' + @Column_Name + ')),'''''''','''''''''''')+'''''''',''NULL'')'
 WHEN @Data_Type IN ('text') 
 THEN 
 'COALESCE('''''''' + REPLACE(CONVERT(varchar(max),' + @Column_Name + '),'''''''','''''''''''')+'''''''',''NULL'')' 
 WHEN @Data_Type IN ('ntext') 
 THEN 
 'COALESCE('''''''' + REPLACE(CONVERT(nvarchar(max),' + @Column_Name + '),'''''''','''''''''''')+'''''''',''NULL'')' 
 WHEN @Data_Type IN ('binary','varbinary') 
 THEN 
 'COALESCE(RTRIM(CONVERT(char,' + 'CONVERT(int,' + @Column_Name + '))),''NULL'')' 
 WHEN @Data_Type IN ('timestamp','rowversion') 
 THEN 
 CASE 
 WHEN @include_timestamp = 0 
 THEN 
 '''DEFAULT''' 
 ELSE 
 'COALESCE(RTRIM(CONVERT(char,' + 'CONVERT(int,' + @Column_Name + '))),''NULL'')' 
 END
 WHEN @Data_Type IN ('float','real','money','smallmoney')
 THEN
 'COALESCE(LTRIM(RTRIM(' + 'CONVERT(char, ' + @Column_Name + ',2)' + ')),''NULL'')' 
 ELSE 
 'COALESCE(LTRIM(RTRIM(' + 'CONVERT(char, ' + @Column_Name + ')' + ')),''NULL'')' 
 END + '+' + ''',''' + ' + '

 --Generating the column list for the MERGE statement
 SET @Column_List = @Column_List + @Column_Name + ',' 

 --Don't update Primary Key or Identity columns
 IF NOT EXISTS(
 SELECT 1
 FROM INFORMATION_SCHEMA.TABLE_CONSTRAINTS pk ,
 INFORMATION_SCHEMA.KEY_COLUMN_USAGE c
 WHERE pk.TABLE_NAME = @table_name
 AND pk.TABLE_SCHEMA = COALESCE(@schema, SCHEMA_NAME())
 AND CONSTRAINT_TYPE = 'PRIMARY KEY'
 AND c.TABLE_NAME = pk.TABLE_NAME
 AND c.TABLE_SCHEMA = pk.TABLE_SCHEMA
 AND c.CONSTRAINT_NAME = pk.CONSTRAINT_NAME
 AND c.COLUMN_NAME = @Column_Name_Unquoted 
 )
 BEGIN
 SET @Column_List_For_Update = @Column_List_For_Update + @Column_Name + ' = Source.' + @Column_Name + ', 
' 
 SET @Column_List_For_Check = @Column_List_For_Check +
 CASE @Data_Type 
 WHEN 'text' THEN CHAR(10) + CHAR(9) + 'NULLIF(CAST(Source.' + @Column_Name + ' AS VARCHAR(MAX)), CAST(Target.' + @Column_Name + ' AS VARCHAR(MAX))) IS NOT NULL OR NULLIF(CAST(Target.' + @Column_Name + ' AS VARCHAR(MAX)), CAST(Source.' + @Column_Name + ' AS VARCHAR(MAX))) IS NOT NULL OR '
 WHEN 'ntext' THEN CHAR(10) + CHAR(9) + 'NULLIF(CAST(Source.' + @Column_Name + ' AS NVARCHAR(MAX)), CAST(Target.' + @Column_Name + ' AS NVARCHAR(MAX))) IS NOT NULL OR NULLIF(CAST(Target.' + @Column_Name + ' AS NVARCHAR(MAX)), CAST(Source.' + @Column_Name + ' AS NVARCHAR(MAX))) IS NOT NULL OR ' 
 ELSE CHAR(10) + CHAR(9) + 'NULLIF(Source.' + @Column_Name + ', Target.' + @Column_Name + ') IS NOT NULL OR NULLIF(Target.' + @Column_Name + ', Source.' + @Column_Name + ') IS NOT NULL OR '
 END 
 END

 SKIP_LOOP: --The label used in GOTO

 SELECT @Column_ID = MIN(ORDINAL_POSITION) 
 FROM INFORMATION_SCHEMA.COLUMNS (NOLOCK) 
 WHERE TABLE_NAME = @table_name
 AND TABLE_SCHEMA = COALESCE(@schema, SCHEMA_NAME())
 AND ORDINAL_POSITION > @Column_ID

 END --Loop ends here!


--To get rid of the extra characters that got concatenated during the last run through the loop
IF LEN(@Column_List_For_Update) <> 0
 BEGIN
 SET @Column_List_For_Update = ' ' + LEFT(@Column_List_For_Update,len(@Column_List_For_Update) - 4)
 END

IF LEN(@Column_List_For_Check) <> 0
 BEGIN
 SET @Column_List_For_Check = LEFT(@Column_List_For_Check,len(@Column_List_For_Check) - 3)
 END

SET @Actual_Values = LEFT(@Actual_Values,len(@Actual_Values) - 6)

SET @Column_List = LEFT(@Column_List,len(@Column_List) - 1)
IF LEN(LTRIM(@Column_List)) = 0
 BEGIN
 RAISERROR('No columns to select. There should at least be one column to generate the output',16,1)
 RETURN -1 --Failure. Reason: Looks like all the columns are ommitted using the @cols_to_exclude parameter
 END


--Get the join columns ----------------------------------------------------------
DECLARE @PK_column_list VARCHAR(8000)
DECLARE @PK_column_joins VARCHAR(8000)
SET @PK_column_list = ''
SET @PK_column_joins = ''

SELECT @PK_column_list = @PK_column_list + c.COLUMN_NAME + ', '
, @PK_column_joins = @PK_column_joins + 'Target.[' + c.COLUMN_NAME + '] = Source.[' + c.COLUMN_NAME + '] AND '
FROM INFORMATION_SCHEMA.TABLE_CONSTRAINTS pk ,
INFORMATION_SCHEMA.KEY_COLUMN_USAGE c
WHERE pk.TABLE_NAME = @table_name
AND pk.TABLE_SCHEMA = COALESCE(@schema, SCHEMA_NAME())
AND CONSTRAINT_TYPE = 'PRIMARY KEY'
AND c.TABLE_NAME = pk.TABLE_NAME
AND c.TABLE_SCHEMA = pk.TABLE_SCHEMA
AND c.CONSTRAINT_NAME = pk.CONSTRAINT_NAME

IF IsNull(@PK_column_list, '') = '' 
 BEGIN
 RAISERROR('Table has no primary keys. There should at least be one column in order to have a valid join.',16,1)
 RETURN -1 --Failure. Reason: looks like table doesn't have any primary keys
 END

SET @PK_column_list = LEFT(@PK_column_list, LEN(@PK_column_list) -1)
SET @PK_column_joins = LEFT(@PK_column_joins, LEN(@PK_column_joins) -4)


--Forming the final string that will be executed, to output the a MERGE statement
SET @Actual_Values = 
 'SELECT ' + 
 CASE WHEN @top IS NULL OR @top < 0 THEN '' ELSE ' TOP ' + LTRIM(STR(@top)) + ' ' END + 
 '''' + 
 ' '' + CASE WHEN ROW_NUMBER() OVER (ORDER BY ' + @PK_column_list + ') = 1 THEN '' '' ELSE '','' END + ''(''+ ' + @Actual_Values + '+'')''' + ' ' + 
 COALESCE(@from,' FROM ' + @Source_Table_Qualified + ' (NOLOCK)')

 DECLARE @output VARCHAR(MAX) = ''
 DECLARE @b CHAR(1) = CHAR(13)

--Determining whether to ouput any debug information
IF @debug_mode =1
 BEGIN
 SET @output += @b + '/*****START OF DEBUG INFORMATION*****'
 SET @output += @b + ''
 SET @output += @b + 'The primary key column list:'
 SET @output += @b + @PK_column_list
 SET @output += @b + ''
 SET @output += @b + 'The INSERT column list:'
 SET @output += @b + @Column_List
 SET @output += @b + ''
 SET @output += @b + 'The UPDATE column list:'
 SET @output += @b + @Column_List_For_Update
 SET @output += @b + ''
 SET @output += @b + 'The SELECT statement executed to generate the MERGE:'
 SET @output += @b + @Actual_Values
 SET @output += @b + ''
 SET @output += @b + '*****END OF DEBUG INFORMATION*****/'
 SET @output += @b + ''
 END

IF (@include_use_db = 1)
BEGIN
    DECLARE @db varchar(120);
    SET @db = 'USE ' + DB_NAME();
    SET @output += @b + @db;
    SET @output += @b + 'GO';
    SET @output += @b + '';
END

SET @output += @b + '--MERGE generated by ''sp_generate_merge'' stored procedure, Version 0.93'
SET @output += @b + '--Originally by Vyas (http://vyaskn.tripod.com): sp_generate_inserts (build 22)'
SET @output += @b + '--Adapted for SQL Server 2008/2012 by Daniel Nolan (http://danere.com)'
SET @output += @b + ''
SET @output += @b + 'SET NOCOUNT ON'
SET @output += @b + ''


--Determining whether to print IDENTITY_INSERT or not
IF (LEN(@IDN) <> 0)
 BEGIN
 SET @output += @b + 'SET IDENTITY_INSERT ' + @Target_Table_For_Output + ' ON'
 SET @output += @b + ''
 END


--Temporarily disable constraints on the target table
IF @disable_constraints = 1 AND (OBJECT_ID(@Source_Table_Qualified, 'U') IS NOT NULL)
 BEGIN
 SET @output += @b + 'ALTER TABLE ' + @Target_Table_For_Output + ' NOCHECK CONSTRAINT ALL' --Code to disable constraints temporarily
 END


--Output the start of the MERGE statement, qualifying with the schema name only if the caller explicitly specified it
SET @output += @b + 'MERGE INTO ' + @Target_Table_For_Output + ' AS Target'
SET @output += @b + 'USING (VALUES'


--All the hard work pays off here!!! You'll get your MERGE statement, when the next line executes!
DECLARE @tab TABLE (val NVARCHAR(max));
INSERT INTO @tab
EXEC (@Actual_Values)

SET @output += CAST((SELECT @b + val FROM @tab FOR XML PATH('')) AS XML).value('.', 'VARCHAR(MAX)');

--Output the columns to correspond with each of the values above--------------------
SET @output += @b + ') AS Source (' + @Column_List + ')'


--Output the join columns ----------------------------------------------------------
SET @output += @b + 'ON (' + @PK_column_joins + ')'


--When matched, perform an UPDATE on any metadata columns only (ie. not on PK)------
IF LEN(@Column_List_For_Update) <> 0
BEGIN
 SET @output += @b + 'WHEN MATCHED ' + CASE WHEN @update_only_if_changed = 1 THEN 'AND (' + @Column_List_For_Check + ') ' ELSE '' END + 'THEN'
 SET @output += @b + ' UPDATE SET'
 SET @output += @b + @Column_List_For_Update
END


--When NOT matched by target, perform an INSERT------------------------------------
SET @output += @b + 'WHEN NOT MATCHED BY TARGET THEN';
SET @output += @b + ' INSERT(' + @Column_List + ')'
SET @output += @b + ' VALUES(' + REPLACE(@Column_List, '[', 'Source.[') + ')'


--When NOT matched by source, DELETE the row
SET @output += @b + 'WHEN NOT MATCHED BY SOURCE THEN '
SET @output += @b + ' DELETE;'
SET @output += @b + ''
SET @output += @b + 'GO'

--Display the number of affected rows to the user, or report if an error occurred---
SET @output += @b + 'DECLARE @mergeError int'
SET @output += @b + ' , @mergeCount int'
SET @output += @b + 'SELECT @mergeError = @@ERROR, @mergeCount = @@ROWCOUNT'
SET @output += @b + 'IF @mergeError != 0'
SET @output += @b + ' BEGIN'
SET @output += @b + ' PRINT ''ERROR OCCURRED IN MERGE FOR ' + @Target_Table_For_Output + '. Rows affected: '' + CAST(@mergeCount AS VARCHAR(100)); -- SQL should always return zero rows affected';
SET @output += @b + ' END'
SET @output += @b + 'ELSE'
SET @output += @b + ' BEGIN'
SET @output += @b + ' PRINT ''' + @Target_Table_For_Output + ' rows affected by MERGE: '' + CAST(@mergeCount AS VARCHAR(100));';
SET @output += @b + ' END'
SET @output += @b + 'GO'
SET @output += @b + ''

--Re-enable the previously disabled constraints-------------------------------------
IF @disable_constraints = 1 AND (OBJECT_ID(@Source_Table_Qualified, 'U') IS NOT NULL)
 BEGIN
 SET @output += @b + 'ALTER TABLE ' + @Target_Table_For_Output + ' CHECK CONSTRAINT ALL' --Code to enable the previously disabled constraints
 SET @output += @b + 'GO'
 END


--Switch-off identity inserting------------------------------------------------------
IF (LEN(@IDN) <> 0)
 BEGIN
 SET @output += @b + 'SET IDENTITY_INSERT ' + @Target_Table_For_Output + ' OFF'
 SET @output += @b + 'GO'
 END


SET @output += @b + 'SET NOCOUNT OFF'
SET @output += @b + 'GO'
SET @output += @b + ''
SET @output += @b + ''

IF @results_to_text = 1
BEGIN
    --output the statement to the Grid/Messages tab
    SELECT @output;
END
ELSE
BEGIN
    --output the statement as xml (to overcome SSMS 4000/8000 char limitation)
SELECT [processing-instruction(x)]=@output FOR XML PATH(''),TYPE;
    PRINT 'MERGE statement has been wrapped in an XML fragment and output successfully.'
    PRINT 'Ensure you have Results to Grid enabled and then click the hyperlink to copy the statement within the fragment.'
    PRINT ''
    PRINT 'If you would prefer to have results output directly (without XML) specify @results_to_text = 1, however please'
    PRINT 'note that the results may be truncated by your SQL client to 4000 nchars.'
END

SET NOCOUNT OFF
RETURN 0 --Success. We are done!
END

GO

PRINT 'Created the procedure'
GO


--Mark the proc as a system object to allow it to be called transparently from other databases
EXEC sp_MS_marksystemobject sp_generate_merge
GO

PRINT 'Granting EXECUTE permission on sp_generate_merge to all users'
GRANT EXEC ON sp_generate_merge TO public

SET NOCOUNT OFF
GO

PRINT 'Done'

用法

  1. 确保您的 SQL 客户端配置为将结果发送到网格。
  2. 执行 proc,提供源表名作为参数
  3. 单击结果集中的超链接。
  4. 复制 SQL(不包括输出标签)并粘贴到新的查询窗口中执行。

例子

要生成包含 Person.AddressType 表中所有数据的 MERGE 语句,不包括 ModifiedDate 和 rowguid 列:

EXEC AdventureWorks.dbo.sp_generate_merge @schema = 'Person', @table_name ='AddressType', @cols_to_exclude = '''ModifiedDate'',''rowguid'''

生成的 SQL:

SET NOCOUNT ON
GO 
SET IDENTITY_INSERT [Person].[AddressType] ON
GO

MERGE INTO [Person].[AddressType] AS Target
USING (VALUES
---------------------------------------------------------------------------------------------------------------
  (1,'Billing')
 ,(2,'Home')
 ,(3,'Main Office')
 ,(4,'Primary')
 ,(5,'Shipping')
 ,(6,'Contact')

) AS Source ([AddressTypeID],[Name])
ON (Target.[AddressTypeID] = Source.[AddressTypeID])
WHEN MATCHED AND (
    NULLIF(Source.[Name], Target.[Name]) IS NOT NULL OR NULLIF(Target.[Name], Source.[Name]) IS NOT NULL) THEN
 UPDATE SET
 [Name] = Source.[Name]
WHEN NOT MATCHED BY TARGET THEN
 INSERT([AddressTypeID],[Name])
 VALUES(Source.[AddressTypeID],Source.[Name])
WHEN NOT MATCHED BY SOURCE THEN 
 DELETE;

SET IDENTITY_INSERT [Person].[AddressType] OFF
GO
SET NOCOUNT OFF
GO
于 2012-05-10T07:48:16.577 回答
2

如果我对您的理解正确,您在一个环境中有一个表(我们称之为 DEV),并且您想在另一个环境中更新(我们称之为 PROD)。出于显而易见的原因,您不允许通过数据库链接连接这两者。

最好的选择可能是

  1. BCP 将 DEV 表的内容保存到文件中
  2. BCP 文件到 PROD 中的临时表中
  3. 编写一个合并语句,从临时表中插入/更新/删除您的值。
  4. 清除/删除临时表

这样做的好处是,如果您定期执行此操作,则很容易重复使用不同的源文件。

当然,如果您可以访问并被允许使用这些工具,SSIS 或 Redgate 数据比较将使这变得更容易。

于 2012-05-07T21:18:36.647 回答