Wie lassen sich INSERT-Anweisungen für eine SQL Server-Tabelle am besten automatisch generieren?

Question

Wir schreiben eine neue Anwendung, und während des Testens benötigen wir eine Reihe von Dummy-Daten. Ich habe diese Daten mithilfe von MS Access hinzugefügt, um Excel-Dateien in die entsprechenden Tabellen zu übertragen.

Von Zeit zu Zeit wollen wir die betreffenden Tabellen "auffrischen", was bedeutet, dass wir sie alle löschen, neu erstellen und eine gespeicherte MS Access-Anhangsabfrage ausführen.

Der erste Teil (Löschen und Neuanlegen) ist ein einfaches SQL-Skript, aber der letzte Teil lässt mich erschaudern. Ich möchte ein einziges Setup-Skript, das eine Reihe von INSERTs hat, um die Dummy-Daten zu regenerieren.

Ich habe die Daten jetzt in den Tabellen. Wie kann ich am besten automatisch eine große Liste von INSERT-Anweisungen aus diesem Datensatz erstellen?

Der einzige Weg, den ich mir vorstellen kann, ist, die Tabelle in einer Excel-Tabelle zu speichern und dann eine Excel-Formel zu schreiben, um für jede Zeile eine EINFÜGE zu erstellen, was sicherlich nicht der beste Weg ist.

Ich verwende das Management Studio 2008, um eine Verbindung zu einer SQL Server 2005-Datenbank herzustellen.

Answer 1

Ich verwende dazu Sqlite. Ich finde es sehr, sehr nützlich für die Erstellung von Scratch-/Testdatenbanken.

sqlite3 foo.sqlite .dump > foo_as_a_bunch_of_inserts.sql

Answer 2

Es gibt viele gute Skripte für die Erstellung von Einfügeanweisungen, aber ich habe versucht, ein eigenes Skript zu erstellen, um es so benutzerfreundlich wie möglich zu machen und um auch UPDATE-Anweisungen ausführen zu können. + Das Ergebnis wird in .sql-Dateien verpackt, die nach Datum gespeichert werden können.

Es nimmt als Eingabe Ihre normale SELECT-Anweisung mit WHERE-Klausel und gibt dann eine Liste von Insert-Anweisungen und Update-Anweisungen aus. Zusammen bilden sie eine Art IF NOT EXISTS () INSERT ELSE UPDATE Es ist auch praktisch, wenn es nicht aktualisierbare Spalten gibt, die von der endgültigen INSERT/UPDATE-Anweisung ausgeschlossen werden müssen.

Eine andere Sache, die das folgende Skript kann, ist: Es kann sogar INNER JOINs mit anderen Tabellen als Eingabeanweisung für den Stored Proc verarbeiten. Das Skript ist ein praktisches Tool für die Freigabeverwaltung, das direkt an Ihren Fingerspitzen sitzt, wenn Sie den ganzen Tag SELECT-Anweisungen eingeben.

Originalbeitrag : Erzeugen einer UPDATE-Anweisung in SQL Server für eine bestimmte Tabelle

CREATE PROCEDURE [dbo].[sp_generate_updates] (
    @fullquery              nvarchar(max) = '',
    @ignore_field_input     nvarchar(MAX) = '',
    @PK_COLUMN_NAME         nvarchar(MAX) = ''
)
AS

SET NOCOUNT ON
SET QUOTED_IDENTIFIER ON
/*
-- For Standard USAGE: (where clause is mandatory)
                EXEC [sp_generate_updates] 'select * from dbo.mytable where mytext=''1''  ' 
        OR
                SET QUOTED_IDENTIFIER OFF 
                EXEC [sp_generate_updates] "select * from dbo.mytable where mytext='1'    "

-- For ignoring specific columns  (to ignore in the UPDATE and INSERT SQL statement) 
                EXEC [sp_generate_updates] 'select * from dbo.mytable where 1=1  ' , 'Column01,Column02'

-- For just updates without insert statement (replace the * )
                EXEC [sp_generate_updates] 'select Column01, Column02 from dbo.mytable where 1=1  ' 

-- For tables without a primary key: construct the key in the third variable
                EXEC [sp_generate_updates] 'select * from dbo.mytable where 1=1  '  ,'','your_chosen_primary_key_Col1,key_Col2'

-- For complex updates with JOINED tables 
                EXEC [sp_generate_updates] 'select o1.Name,  o1.category, o2.name+ '_hello_world' as #name 
                                            from overnightsetting o1 
                                            inner join overnightsetting o2  on o1.name=o2.name  
                                            where o1.name like '%appserver%' 
                (REMARK about above:   the use of # in front of a column name (so #abc) can do an update of that columname (abc) with any column from an inner joined table where you use the alias #abc )

-------------README for the deeper interested person:
            Goal of the Stored PROCEDURE is to get updates from simple SQL SELECT statements. It is made ot be simple but fast and powerfull. As always => power is nothing without control, so check before you execute.
            Its power sits also in the fact that you can make insert statements, so combined gives you a  "IF NOT EXISTS()  INSERT "   capability. 

            The scripts work were there are primary keys or identity columns on table you want to update (/ or make inserts for).
            It will also works when no primary keys / identity column exist(s) and you define them yourselve. But then be carefull (duplicate hits can occur). When the table has a primary key it will always be used.
            The script works with a real  temporary table, made on the fly   (APPROPRIATE RIGHTS needed), to put the values inside from the script, then add 3 columns for constructing the "insert into tableX (...) values ()" ,  and the 2 update statement.
            We work with temporary structures like   "where columnname = {Columnname}" and then later do the update on that temptable for the columns values found on that same line.
                    example  "where columnname = {Columnname}"  for birthdate becomes   "where birthdate = {birthdate}" an then we find the birthdate value on that line inside the temp table.
            So then the statement becomes  "where birthdate = {19800417}"
            Enjoy releasing scripts as of now...                                        by  Pieter van Nederkassel  - freeware "CC BY-SA" (+use at own risk)
*/
IF OBJECT_ID('tempdb..#ignore','U') IS NOT NULL     DROP TABLE #ignore
DECLARE @stringsplit_table               TABLE (col nvarchar(255), dtype  nvarchar(255)) -- table to store the primary keys or identity key
DECLARE @PK_condition                    nvarchar(512), -- placeholder for WHERE pk_field1 = pk_value1 AND pk_field2 = pk_value2 AND ...
        @pkstring                        NVARCHAR(512),  -- sting to store the primary keys or the idendity key
        @table_name                      nvarchar(512), -- (left) table name, including schema
        @table_N_where_clause            nvarchar(max), -- tablename 
        @table_alias                     nvarchar(512), -- holds the (left) table alias if one available, else @table_name
        @table_schema                    NVARCHAR(30),  -- schema of @table_name
        @update_list1                    NVARCHAR(MAX), -- placeholder for SET fields section of update
        @update_list2                    NVARCHAR(MAX), -- placeholder for SET fields section of update value comming from other tables in the join, other than the main table to update => updateof base table possible with inner join
        @list_all_cols                   BIT = 0,       -- placeholder for values for the insert into table VALUES command
        @select_list                     NVARCHAR(MAX), -- placeholder for SELECT fields of (left) table
        @COLUMN_NAME                     NVARCHAR(255), -- will hold column names of the (left) table
        @sql                             NVARCHAR(MAX), -- sql statement variable
        @getdate                         NVARCHAR(17),  -- transform getdate() to YYYYMMDDHHMMSSMMM
        @tmp_table                       NVARCHAR(255), -- will hold the name of a physical temp table
        @pk_separator                    NVARCHAR(1),   -- separator used in @PK_COLUMN_NAME if provided (only checking obvious ones ,;|-)
        @COLUMN_NAME_DATA_TYPE           NVARCHAR(100), -- needed for insert statements to convert to right text string
        @own_pk                          BIT = 0        -- check if table has PK (0) or if provided PK will be used (1)

set @ignore_field_input=replace(replace(replace(@ignore_field_input,' ',''),'[',''),']','')
set @PK_COLUMN_NAME=    replace(replace(replace(@PK_COLUMN_NAME,    ' ',''),'[',''),']','')

-- first we remove all linefeeds from the user query
set @fullquery=replace(replace(replace(@fullquery,char(10),''),char(13),' '),'  ',' ')
set @table_N_where_clause=@fullquery
if charindex ('order by' , @table_N_where_clause) > 0
    print ' WARNING:        ORDER BY NOT ALLOWED IN UPDATE ...'
if @PK_COLUMN_NAME <> ''
    select ' WARNING:        IF you select your own primary keys, make double sure before doing the update statements below!! '
--print @table_N_where_clause
if charindex ('select ' , @table_N_where_clause) = 0
    set @table_N_where_clause= 'select * from ' + @table_N_where_clause
if charindex ('select ' , @table_N_where_clause) > 0
    exec (@table_N_where_clause)

set @table_N_where_clause=rtrim(ltrim(substring(@table_N_where_clause,CHARINDEX(' from ', @table_N_where_clause )+6, 4000)))
--print @table_N_where_clause 
set @table_name=left(@table_N_where_clause,CHARINDEX(' ', @table_N_where_clause )-1)

IF CHARINDEX('where ', @table_N_where_clause) > 0             SELECT @table_alias = LTRIM(RTRIM(REPLACE(REPLACE(SUBSTRING(@table_N_where_clause,1, CHARINDEX('where ', @table_N_where_clause )-1),'(nolock)',''),@table_name,'')))
IF CHARINDEX('join ',  @table_alias) > 0                      SELECT @table_alias = SUBSTRING(@table_alias, 1, CHARINDEX(' ', @table_alias)-1) -- until next space
IF LEN(@table_alias) = 0                                      SELECT @table_alias = @table_name
IF (charindex (' *' , @fullquery) > 0 or charindex (@table_alias+'.*' , @fullquery) > 0 )     set @list_all_cols=1
/*       
       print @fullquery     
       print @table_alias
       print @table_N_where_clause
       print @table_name
*/

-- Prepare PK condition
        SELECT @table_schema = CASE WHEN CHARINDEX('.',@table_name) > 0 THEN LEFT(@table_name, CHARINDEX('.',@table_name)-1) ELSE 'dbo' END

        SELECT @PK_condition = ISNULL(@PK_condition + ' AND ', '') + QUOTENAME('pk_'+COLUMN_NAME) + ' = ' + QUOTENAME('pk_'+COLUMN_NAME,'{')
        FROM INFORMATION_SCHEMA.KEY_COLUMN_USAGE
        WHERE OBJECTPROPERTY(OBJECT_ID(CONSTRAINT_SCHEMA + '.' + QUOTENAME(CONSTRAINT_NAME)), 'IsPrimaryKey') = 1
        AND TABLE_NAME = REPLACE(@table_name,@table_schema+'.','') 
        AND TABLE_SCHEMA = @table_schema

        SELECT @pkstring = ISNULL(@pkstring + ', ', '') + @table_alias + '.' + QUOTENAME(COLUMN_NAME) + ' AS pk_' + COLUMN_NAME
        FROM INFORMATION_SCHEMA.KEY_COLUMN_USAGE i1
        WHERE OBJECTPROPERTY(OBJECT_ID(i1.CONSTRAINT_SCHEMA + '.' + QUOTENAME(i1.CONSTRAINT_NAME)), 'IsPrimaryKey') = 1
        AND i1.TABLE_NAME = REPLACE(@table_name,@table_schema+'.','') 
        AND i1.TABLE_SCHEMA = @table_schema

            -- if no primary keys exist then we try for identity columns
                IF @PK_condition is null SELECT @PK_condition = ISNULL(@PK_condition + ' AND ', '') + QUOTENAME('pk_'+COLUMN_NAME) + ' = ' + QUOTENAME('pk_'+COLUMN_NAME,'{')
                FROM  INFORMATION_SCHEMA.COLUMNS
                WHERE COLUMNPROPERTY(object_id(TABLE_SCHEMA+'.'+TABLE_NAME), COLUMN_NAME, 'IsIdentity') = 1 
                AND TABLE_NAME = REPLACE(@table_name,@table_schema+'.','') 
                AND TABLE_SCHEMA = @table_schema

                IF @pkstring is null SELECT @pkstring = ISNULL(@pkstring + ', ', '') + @table_alias + '.' + QUOTENAME(COLUMN_NAME) + ' AS pk_' + COLUMN_NAME
                FROM  INFORMATION_SCHEMA.COLUMNS
                WHERE COLUMNPROPERTY(object_id(TABLE_SCHEMA+'.'+TABLE_NAME), COLUMN_NAME, 'IsIdentity') = 1 
                AND TABLE_NAME = REPLACE(@table_name,@table_schema+'.','') 
                AND TABLE_SCHEMA = @table_schema
-- Same but in form of a table

        INSERT INTO @stringsplit_table
        SELECT 'pk_'+i1.COLUMN_NAME as col, i2.DATA_TYPE as dtype
        FROM INFORMATION_SCHEMA.KEY_COLUMN_USAGE i1
        inner join INFORMATION_SCHEMA.COLUMNS i2
        on  i1.TABLE_NAME = i2.TABLE_NAME AND i1.TABLE_SCHEMA =  i2.TABLE_SCHEMA
        WHERE OBJECTPROPERTY(OBJECT_ID(i1.CONSTRAINT_SCHEMA + '.' + QUOTENAME(i1.CONSTRAINT_NAME)), 'IsPrimaryKey') = 1
        AND i1.TABLE_NAME = REPLACE(@table_name,@table_schema+'.','') 
        AND i1.TABLE_SCHEMA = @table_schema

                -- if no primary keys exist then we try for identity columns
                IF 0=(select count(*) from @stringsplit_table) INSERT INTO @stringsplit_table
                SELECT 'pk_'+i2.COLUMN_NAME as col, i2.DATA_TYPE as dtype
                FROM INFORMATION_SCHEMA.COLUMNS i2
                WHERE COLUMNPROPERTY(object_id(i2.TABLE_SCHEMA+'.'+i2.TABLE_NAME), i2.COLUMN_NAME, 'IsIdentity') = 1 
                AND i2.TABLE_NAME = REPLACE(@table_name,@table_schema+'.','') 
                AND i2.TABLE_SCHEMA = @table_schema

-- NOW handling the primary key given as parameter to the main batch

SELECT @pk_separator = ',' -- take this as default, we'll check lower if it's a different one
IF (@PK_condition IS NULL OR @PK_condition = '') AND @PK_COLUMN_NAME <> ''
BEGIN
    IF CHARINDEX(';', @PK_COLUMN_NAME) > 0
        SELECT @pk_separator = ';'
    ELSE IF CHARINDEX('|', @PK_COLUMN_NAME) > 0
        SELECT @pk_separator = '|'
    ELSE IF CHARINDEX('-', @PK_COLUMN_NAME) > 0
        SELECT @pk_separator = '-'

    SELECT @PK_condition = NULL -- make sure to make it NULL, in case it was ''
    INSERT INTO @stringsplit_table
    SELECT LTRIM(RTRIM(x.value)) , 'datetime'  FROM STRING_SPLIT(@PK_COLUMN_NAME, @pk_separator) x  
    SELECT @PK_condition = ISNULL(@PK_condition + ' AND ', '') + QUOTENAME(x.col) + ' = ' + replace(QUOTENAME(x.col,'{'),'{','{pk_')
      FROM @stringsplit_table x

    SELECT @PK_COLUMN_NAME = NULL -- make sure to make it NULL, in case it was ''
    SELECT @PK_COLUMN_NAME = ISNULL(@PK_COLUMN_NAME + ', ', '') + QUOTENAME(x.col) + ' as pk_' + x.col
      FROM @stringsplit_table x
    --print 'pkcolumns  '+ isnull(@PK_COLUMN_NAME,'')
    update @stringsplit_table set col='pk_' + col
    SELECT @own_pk = 1
END
ELSE IF (@PK_condition IS NULL OR @PK_condition = '') AND @PK_COLUMN_NAME = ''
BEGIN
    RAISERROR('No Primary key or Identity column available on table. Add some columns as the third parameter when calling this SP to make your own temporary PK., also remove  [] from tablename',17,1)
END

-- IF there are no primary keys or an identity key in the table active, then use the given columns as a primary key

if isnull(@pkstring,'')   = ''  set    @pkstring  = @PK_COLUMN_NAME
IF ISNULL(@pkstring, '') <> ''  SELECT @fullquery = REPLACE(@fullquery, 'SELECT ','SELECT ' + @pkstring + ',' )
--print @pkstring

-- ignore fields for UPDATE STATEMENT (not ignored for the insert statement,  in iserts statement we ignore only identity Columns and the columns provided with the main stored proc )
-- Place here all fields that you know can not be converted to nvarchar() values correctly, an thus should not be scripted for updates)
-- for insert we will take these fields along, although they will be incorrectly represented!!!!!!!!!!!!!.
SELECT           ignore_field = 'uniqueidXXXX' INTO #ignore 
UNION ALL SELECT ignore_field = 'UPDATEMASKXXXX'
UNION ALL SELECT ignore_field = 'UIDXXXXX'
UNION ALL SELECT value FROM  string_split(@ignore_field_input,@pk_separator)

SELECT @getdate = REPLACE(REPLACE(REPLACE(REPLACE(CONVERT(NVARCHAR(30), GETDATE(), 121), '-', ''), ' ', ''), ':', ''), '.', '')
SELECT @tmp_table = 'Release_DATA__' + @getdate + '__' + REPLACE(@table_name,@table_schema+'.','') 

SET @sql = replace( @fullquery,  ' from ',  ' INTO ' + @tmp_table +' from ')
----print (@sql)
exec (@sql)

SELECT @sql = N'alter table ' + @tmp_table + N'  add update_stmt1  nvarchar(max), update_stmt2 nvarchar(max) , update_stmt3 nvarchar(max)'
EXEC (@sql)

-- Prepare update field list (only columns from the temp table are taken if they also exist in the base table to update)
SELECT @update_list1 = ISNULL(@update_list1 + ', ', '') + 
                      CASE WHEN C1.COLUMN_NAME = 'ModifiedBy' THEN '[ModifiedBy] = left(right(replace(CONVERT(VARCHAR(19),[Modified],121),''''-'''',''''''''),19) +''''-''''+right(SUSER_NAME(),30),50)'
                           WHEN C1.COLUMN_NAME = 'Modified' THEN '[Modified] = GETDATE()'
                           ELSE QUOTENAME(C1.COLUMN_NAME) + ' = ' + QUOTENAME(C1.COLUMN_NAME,'{')
                      END
FROM INFORMATION_SCHEMA.COLUMNS c1
inner join INFORMATION_SCHEMA.COLUMNS c2
on c1.COLUMN_NAME =c2.COLUMN_NAME and c2.TABLE_NAME = REPLACE(@table_name,@table_schema+'.','')  AND c2.TABLE_SCHEMA = @table_schema
WHERE c1.TABLE_NAME = @tmp_table --REPLACE(@table_name,@table_schema+'.','') 
AND QUOTENAME(c1.COLUMN_NAME) NOT IN (SELECT QUOTENAME(ignore_field) FROM #ignore) -- eliminate binary, image etc value here
AND COLUMNPROPERTY(object_id(c2.TABLE_SCHEMA+'.'+c2.TABLE_NAME), c2.COLUMN_NAME, 'IsIdentity') <> 1
AND NOT EXISTS (SELECT 1 
                  FROM INFORMATION_SCHEMA.KEY_COLUMN_USAGE ku 
                 WHERE 1 = 1
                   AND ku.TABLE_NAME = c2.TABLE_NAME
                   AND ku.TABLE_SCHEMA = c2.TABLE_SCHEMA
                   AND ku.COLUMN_NAME = c2.COLUMN_NAME
                   AND OBJECTPROPERTY(OBJECT_ID(ku.CONSTRAINT_SCHEMA + '.' + QUOTENAME(ku.CONSTRAINT_NAME)), 'IsPrimaryKey') = 1)
AND NOT EXISTS (SELECT 1 FROM @stringsplit_table x WHERE x.col = c2.COLUMN_NAME AND @own_pk = 1)

-- Prepare update field list  (here we only take columns that commence with a #, as this is our queue for doing the update that comes from an inner joined table)
SELECT @update_list2 = ISNULL(@update_list2 + ', ', '') +  QUOTENAME(replace( C1.COLUMN_NAME,'#','')) + ' = ' + QUOTENAME(C1.COLUMN_NAME,'{')
FROM INFORMATION_SCHEMA.COLUMNS c1
WHERE c1.TABLE_NAME = @tmp_table --AND c1.TABLE_SCHEMA = @table_schema
AND QUOTENAME(c1.COLUMN_NAME) NOT IN (SELECT QUOTENAME(ignore_field) FROM #ignore) -- eliminate binary, image etc value here
AND c1.COLUMN_NAME like '#%'

-- similar for select list, but take all fields
SELECT @select_list = ISNULL(@select_list + ', ', '') + QUOTENAME(COLUMN_NAME)
FROM INFORMATION_SCHEMA.COLUMNS c
WHERE TABLE_NAME = REPLACE(@table_name,@table_schema+'.','') 
AND TABLE_SCHEMA = @table_schema
AND COLUMNPROPERTY(object_id(TABLE_SCHEMA+'.'+TABLE_NAME), COLUMN_NAME, 'IsIdentity') <> 1  -- Identity columns are filled automatically by MSSQL, not needed at Insert statement
AND QUOTENAME(c.COLUMN_NAME) NOT IN (SELECT QUOTENAME(ignore_field) FROM #ignore) -- eliminate binary, image etc value here

SELECT @PK_condition = REPLACE(@PK_condition, '[pk_', '[')
set @select_list='if not exists (select * from '+  REPLACE(@table_name,@table_schema+'.','') +'  where '+  @PK_condition +')  INSERT INTO '+ REPLACE(@table_name,@table_schema+'.','')   + '('+ @select_list  + ') VALUES (' + replace(replace(@select_list,'[','{'),']','}') + ')'
SELECT @sql = N'UPDATE ' + @tmp_table + ' set update_stmt1 = ''' + @select_list + '''' 
if @list_all_cols=1 EXEC (@sql)

--print 'select==========  ' + @select_list
--print 'update==========  ' + @update_list1

SELECT @sql = N'UPDATE ' + @tmp_table + N'
set update_stmt2 = CONVERT(NVARCHAR(MAX),''UPDATE ' + @table_name + 
                                          N' SET ' + @update_list1 + N''' + ''' +
                                          N' WHERE ' + @PK_condition + N''') ' 

EXEC (@sql)
--print @sql

SELECT @sql = N'UPDATE ' + @tmp_table + N'
set update_stmt3 = CONVERT(NVARCHAR(MAX),''UPDATE ' + @table_name + 
                                          N' SET ' + @update_list2 + N''' + ''' +
                                          N' WHERE ' + @PK_condition + N''') ' 

EXEC (@sql)
--print @sql

-- LOOPING OVER ALL base tables column for the INSERT INTO .... VALUES
DECLARE c_columns CURSOR FAST_FORWARD READ_ONLY FOR
    SELECT COLUMN_NAME, DATA_TYPE
    FROM INFORMATION_SCHEMA.COLUMNS 
    WHERE TABLE_NAME = (CASE WHEN @list_all_cols=0 THEN @tmp_table ELSE REPLACE(@table_name,@table_schema+'.','') END )
    AND TABLE_SCHEMA = @table_schema
        UNION--pned
    SELECT col, 'datetime' FROM @stringsplit_table

OPEN c_columns
FETCH NEXT FROM c_columns INTO @COLUMN_NAME, @COLUMN_NAME_DATA_TYPE
WHILE @@FETCH_STATUS = 0
BEGIN
    SELECT @sql = 
    CASE WHEN @COLUMN_NAME_DATA_TYPE IN ('char','varchar','nchar','nvarchar')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt1 = REPLACE(update_stmt1, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL'')) ' 
        WHEN @COLUMN_NAME_DATA_TYPE IN ('float','real','money','smallmoney')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt1 = REPLACE(update_stmt1, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'],126)),   '''''''','''''''''''') + '''''''', ''NULL'')) '
        WHEN @COLUMN_NAME_DATA_TYPE IN ('uniqueidentifier')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt1 = REPLACE(update_stmt1, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL'')) ' 
        WHEN @COLUMN_NAME_DATA_TYPE IN ('text','ntext')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt1 = REPLACE(update_stmt1, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL'')) ' 
        WHEN @COLUMN_NAME_DATA_TYPE IN ('xxxx','yyyy')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt1 = REPLACE(update_stmt1, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL'')) ' 
        WHEN @COLUMN_NAME_DATA_TYPE IN ('binary','varbinary')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt1 = REPLACE(update_stmt1, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL'')) ' 
        WHEN @COLUMN_NAME_DATA_TYPE IN ('XML','xml')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt1 = REPLACE(update_stmt1, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'],0)),     '''''''','''''''''''') + '''''''', ''NULL'')) ' 
        WHEN @COLUMN_NAME_DATA_TYPE IN ('datetime','smalldatetime')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt1 = REPLACE(update_stmt1, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'],121)),   '''''''','''''''''''') + '''''''', ''NULL'')) '
    ELSE  
                  N'UPDATE ' + @tmp_table + N' SET update_stmt1 = REPLACE(update_stmt1, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL'')) '
    END
    ----PRINT @sql
    EXEC (@sql)
    FETCH NEXT FROM c_columns INTO @COLUMN_NAME, @COLUMN_NAME_DATA_TYPE
END
CLOSE c_columns
DEALLOCATE c_columns

--SELECT col FROM @stringsplit_table -- these are the primary keys

-- LOOPING OVER ALL temp tables column for the Update values
DECLARE c_columns CURSOR FAST_FORWARD READ_ONLY FOR
    SELECT COLUMN_NAME,DATA_TYPE
    FROM INFORMATION_SCHEMA.COLUMNS 
    WHERE TABLE_NAME =  @tmp_table --    AND TABLE_SCHEMA = @table_schema
       UNION--pned
    SELECT col, 'datetime' FROM @stringsplit_table

OPEN c_columns
FETCH NEXT FROM c_columns INTO @COLUMN_NAME, @COLUMN_NAME_DATA_TYPE
WHILE @@FETCH_STATUS = 0
BEGIN
    SELECT @sql = 
    CASE WHEN @COLUMN_NAME_DATA_TYPE IN ('char','varchar','nchar','nvarchar')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt2 = REPLACE(update_stmt2, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL'')), update_stmt3 = REPLACE(update_stmt3, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL'')) ' 
        WHEN @COLUMN_NAME_DATA_TYPE IN ('float','real','money','smallmoney')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt2 = REPLACE(update_stmt2, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'],126)),   '''''''','''''''''''') + '''''''', ''NULL'')), update_stmt3 = REPLACE(update_stmt3, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'],126)),   '''''''','''''''''''') + '''''''', ''NULL'')) '
        WHEN @COLUMN_NAME_DATA_TYPE IN ('uniqueidentifier')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt2 = REPLACE(update_stmt2, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL'')), update_stmt3 = REPLACE(update_stmt3, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL''))  ' 
        WHEN @COLUMN_NAME_DATA_TYPE IN ('text','ntext')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt2 = REPLACE(update_stmt2, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL'')), update_stmt3 = REPLACE(update_stmt3, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL''))  ' 
        WHEN @COLUMN_NAME_DATA_TYPE IN ('xxxx','yyyy')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt2 = REPLACE(update_stmt2, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL'')), update_stmt3 = REPLACE(update_stmt3, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL''))  ' 
        WHEN @COLUMN_NAME_DATA_TYPE IN ('binary','varbinary')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt2 = REPLACE(update_stmt2, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL'')), update_stmt3 = REPLACE(update_stmt3, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL''))  ' 
        WHEN @COLUMN_NAME_DATA_TYPE IN ('XML','xml')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt2 = REPLACE(update_stmt2, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'],0)),     '''''''','''''''''''') + '''''''', ''NULL'')), update_stmt3 = REPLACE(update_stmt3, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'],0)),     '''''''','''''''''''') + '''''''', ''NULL''))  ' 
        WHEN @COLUMN_NAME_DATA_TYPE IN ('datetime','smalldatetime')
            THEN  N'UPDATE ' + @tmp_table + N' SET update_stmt2 = REPLACE(update_stmt2, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'],121)),   '''''''','''''''''''') + '''''''', ''NULL'')), update_stmt3 = REPLACE(update_stmt3, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'],121)),   '''''''','''''''''''') + '''''''', ''NULL''))  ' 
    ELSE    
                  N'UPDATE ' + @tmp_table + N' SET update_stmt2 = REPLACE(update_stmt2, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL'')), update_stmt3 = REPLACE(update_stmt3, ''{' + @COLUMN_NAME + N'}'', ISNULL('''''''' + REPLACE(RTRIM(CONVERT(NVARCHAR(MAX),[' + @COLUMN_NAME + N'])),       '''''''','''''''''''') + '''''''', ''NULL''))  ' 
    END
    EXEC (@sql)
    ----print @sql
    FETCH NEXT FROM c_columns INTO @COLUMN_NAME, @COLUMN_NAME_DATA_TYPE
END
CLOSE c_columns
DEALLOCATE c_columns

SET @sql = 'Select * from  ' + @tmp_table + ';'
--exec (@sql)

SELECT @sql = N'
IF OBJECT_ID(''' + @tmp_table + N''', ''U'') IS NOT NULL
BEGIN
       SELECT   ''USE ' + DB_NAME()  + '''  as executelist 
              UNION ALL
       SELECT   ''GO ''  as executelist 
              UNION ALL
       SELECT   '' /*PRESCRIPT CHECK  */              ' + replace(@fullquery,'''','''''')+''' as executelist 
              UNION ALL
       SELECT update_stmt1 as executelist FROM ' + @tmp_table + N' where update_stmt1 is not null
              UNION ALL
       SELECT update_stmt2 as executelist FROM ' + @tmp_table + N' where update_stmt2 is not null
              UNION ALL
       SELECT isnull(update_stmt3, '' add more columns inn query please'')  as executelist FROM ' + @tmp_table + N' where update_stmt3 is not null
              UNION ALL
       SELECT ''--EXEC usp_AddInstalledScript 5, 5, 1, 1, 1, ''''' + @tmp_table + '.sql'''', 2 ''  as executelist
              UNION ALL 
       SELECT   '' /*VERIFY WITH:  */              ' + replace(@fullquery,'''','''''')+''' as executelist 
              UNION ALL
       SELECT ''-- SCRIPT LOCATION:      F:\CopyPaste\++Distributionpoint++\Release_Management\' + @tmp_table + '.sql''  as executelist   
END'
exec (@sql)

SET @sql = 'DROP TABLE ' + @tmp_table + ';'
exec (@sql)

Answer 3

Warum sichern Sie nicht einfach die Daten, bevor Sie mit ihnen arbeiten, und stellen sie wieder her, wenn Sie sie aktualisieren wollen?

wenn Sie Einschübe erzeugen müssen, versuchen Sie es: http://vyaskn.tripod.com/code.htm#inserts

Answer 4

Wenn Sie lieber Google Sheets verwenden möchten, benutzen Sie SeekWell um die Tabelle an ein Blatt zu senden und dann Zeilen nach einem Zeitplan einzufügen, wenn sie dem Blatt hinzugefügt werden.

Siehe aquí für den schrittweisen Ablauf, oder sehen Sie sich einen Video-Demo der Funktion hier.

Answer 5

Haben Sie bereits Daten in einer Produktionsdatenbank? Wenn ja, könnten Sie eine periodische Aktualisierung der Daten über DTS einrichten. Wir aktualisieren unsere Daten wöchentlich an den Wochenenden, und es ist sehr angenehm, jede Woche saubere, echte Daten für unsere Tests zu haben.

Wenn Sie noch nicht in Produktion sind, sollten Sie eine Datenbank erstellen, die Ihren Vorstellungen entspricht (frisch). Dann duplizieren Sie diese Datenbank und verwenden diese neu erstellte Datenbank als Testumgebung. Wenn Sie die saubere Version benötigen, duplizieren Sie einfach die saubere Version erneut und Bob ist dein Onkel .