Need Create View Script for all tables

  • A simpler way to avoid "Select * " is to use "for xml path" to get columns:

    SELECT 'CREATE VIEW ' + TABLE_SCHEMA + '.vw_' + TABLE_NAME + ' AS SELECT '

    + STUFF((

    SELECT ',' + COLUMN_NAME

    FROM INFORMATION_SCHEMA.COLUMNS

    WHERE TABLE_NAME=isc.TABLE_NAME

    FOR XML PATH('')),1,1,'')

    + ' FROM '

    + TABLE_SCHEMA +'.'+ TABLE_NAME + ' SCHEMA_BINDING'

    FROM INFORMATION_SCHEMA.TABLES isc

    where TABLE_TYPE='BASE TABLE'

  • andrew.mckee (6/17/2013)


    A simpler way to avoid "Select * " is to use "for xml path" to get columns:

    SELECT 'CREATE VIEW ' + TABLE_SCHEMA + '.vw_' + TABLE_NAME + ' AS SELECT '

    + STUFF((

    SELECT ',' + COLUMN_NAME

    FROM INFORMATION_SCHEMA.COLUMNS

    WHERE TABLE_NAME=isc.TABLE_NAME

    FOR XML PATH('')),1,1,'')

    + ' FROM '

    + TABLE_SCHEMA +'.'+ TABLE_NAME + ' SCHEMA_BINDING'

    FROM INFORMATION_SCHEMA.TABLES isc

    where TABLE_TYPE='BASE TABLE'

    Without getting in to whether or not view generation is a good idea for the reasons mentioned in the original post (e.g. SSRS, locking, security etc.) I believe that this last post is the best answer.

    Naming the columns is the safest bet.

    If you want to keep the views then adding a little formatting wouldn't go amiss [use char(10)] and I am not sure whether or not it might be a good idea to add an order by to the column list. It might be implied already depending on how INFORMATION_SCHEMA orders itself but it cannot hurt to enforce it.

    SELECT 'CREATE VIEW ' + TABLE_SCHEMA + '.vw_' + TABLE_NAME + ' AS'+ CHAR(10)

    +'SELECT ' + CHAR(10)

    + STUFF((

    SELECT ',' + COLUMN_NAME + CHAR(10)

    FROM INFORMATION_SCHEMA.COLUMNS

    WHERE TABLE_NAME=isc.TABLE_NAME

    ORDER BY ORDINAL_POSITION

    FOR XML PATH('')),1,1,'')

    + ' FROM '

    + TABLE_SCHEMA +'.'+ TABLE_NAME + ' SCHEMA_BINDING' + CHAR(10) + 'GO' + CHAR(10)

    FROM INFORMATION_SCHEMA.TABLES isc

    where TABLE_TYPE='BASE TABLE'

    David Bridge
    David Bridge Technology Limited
    www.davidbridgetechnology.com

  • David,

    Note that this is a 2 year old thread. That said, your code can be improved to avoid errors if table or column names have spaces or other characters. You also missed the schema comparison to prevent problems with tables with the same names on different schemas.

    Here's a modification of your code.

    SELECT 'CREATE VIEW ' + QUOTENAME( Table_Schema ) + '.' + QUOTENAME( 'vw_' + Table_Name )

    + ' AS' + CHAR( 10 )

    + 'SELECT ' + CHAR( 10 )

    + STUFF((

    SELECT ' ,' + QUOTENAME( C.Column_Name ) + CHAR( 10 )

    FROM Information_Schema.Columns C

    WHERE C.Table_Name = Isc.Table_Name

    AND C.Table_Schema = Isc.Table_Schema

    ORDER BY Ordinal_Position

    FOR XML PATH( '' )), 1, 5, ' ' )

    + ' FROM ' + QUOTENAME( Table_Schema ) + '.' + QUOTENAME( Table_Name )

    + ' SCHEMA_BINDING ' + CHAR( 10 )

    + 'GO' + CHAR( 10 )

    FROM Information_Schema.Tables Isc

    WHERE Table_Type = 'BASE TABLE';

    Luis C.
    General Disclaimer:
    Are you seriously taking the advice and code from someone from the internet without testing it? Do you at least understand it? Or can it easily kill your server?

    How to post data/code on a forum to get the best help: Option 1 / Option 2
  • In case anyone thinks this example is a good idea I just want to add my vote to what Sean already suggested, namely: Never, never use SELECT * FROM a table in a view. Using SELECT * like that is a sure route to a whole lot of problems.

    + a billion

    Don Simpson



    I'm not sure about Heisenberg.

  • DonlSimpson (10/28/2014)


    In case anyone thinks this example is a good idea I just want to add my vote to what Sean already suggested, namely: Never, never use SELECT * FROM a table in a view. Using SELECT * like that is a sure route to a whole lot of problems.

    + a billion

    I have to say that "It Depends". Yes, I understand that a view can return the wrong information for a column if SELECT * is used and an "intermediate" column is added but for really wide views (especially from a single table), SELECT * can provide a fair bit of performance difference compared to listing, say, 100 columns. The way to prevent the problem of someone missing the "column shift" problem if someone changes the underlying table is to create the table using WITH SCHEMABINDING to force someone to do something with the view before they change the table.

    --Jeff Moden


    RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
    First step towards the paradigm shift of writing Set Based code:
    ________Stop thinking about what you want to do to a ROW... think, instead, of what you want to do to a COLUMN.

    Change is inevitable... Change for the better is not.


    Helpful Links:
    How to post code problems
    How to Post Performance Problems
    Create a Tally Function (fnTally)

  • Jeff Moden (10/28/2014)


    [...]The way to prevent the problem of someone missing the "column shift" problem if someone changes the underlying table is to create the table using WITH SCHEMABINDING to force someone to do something with the view before they change the table.

    ...which prevents SELECT * as a handy side-effect. If you try:

    Msg 1054, Level 15, State 6, Procedure X, Line 1

    Syntax '*' is not allowed in schema-bound objects.

  • Paul White (10/28/2014)


    Jeff Moden (10/28/2014)


    [...]The way to prevent the problem of someone missing the "column shift" problem if someone changes the underlying table is to create the table using WITH SCHEMABINDING to force someone to do something with the view before they change the table.

    ...which prevents SELECT * as a handy side-effect. If you try:

    Msg 1054, Level 15, State 6, Procedure X, Line 1

    Syntax '*' is not allowed in schema-bound objects.

    Heh... gotta love it. Obviously, I don't use views much. Thank you for the correction, Paul.

    --Jeff Moden


    RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
    First step towards the paradigm shift of writing Set Based code:
    ________Stop thinking about what you want to do to a ROW... think, instead, of what you want to do to a COLUMN.

    Change is inevitable... Change for the better is not.


    Helpful Links:
    How to post code problems
    How to Post Performance Problems
    Create a Tally Function (fnTally)

  • After reading all of the posts here, I put together this script which combines the best of the posts and adds a little more. This script is for making views in a "Destination" DB using tables in a "Source" DB. It needs to be run in the context of the Destination DB.

    It makes views out of all of the tables in the specified SourceDB & SourceSchema.

    declare @SourceDB nvarchar(max), @SourceSchema nvarchar(max), @DestSchema nvarchar(max);

    set @SourceDB = 'MySourceDB';

    set @SourceSchema = 'mySourceSchema';

    set @DestSchema = 'myDestSchema';

    declare @TableInfo table

    (

    Table_Schema nvarchar(max),

    Table_Name nvarchar(max),

    Column_Name nvarchar(max),

    Ordinal_Position int

    );

    declare @sql1 nvarchar(max);

    set @sql1 =

    'select c.Table_Schema, c.Table_Name, c.Column_Name, c.Ordinal_Position

    from ' +

    QUOTENAME(@SourceDB) + '.INFORMATION_SCHEMA.TABLES t inner join ' + QUOTENAME(@SourceDB) + '.INFORMATION_SCHEMA.COLUMNS c on

    t.Table_Schema = c.Table_Schema and

    t.Table_Name = c.Table_Name

    where t.Table_Type = ''BASE TABLE''' +

    case when @SourceSchema is not null then ' and c.Table_Schema = ''' + @SourceSchema + '''' else '' end +

    'group by c.Table_Schema, c.Table_Name, c.Column_Name, c.Ordinal_Position;'

    ;

    insert into @TableInfo

    exec sp_executesql @sql1;

    declare @CreateViewStatements table

    (

    Id int identity(1,1),

    Sql nvarchar(max)

    )

    insert into @CreateViewStatements (Sql)

    SELECT

    'IF EXISTS (SELECT * FROM sys.views

    WHERE object_id = OBJECT_ID('''+ isnull(@DestSchema, TABLE_SCHEMA) +'.view_'+TABLE_NAME+'''))

    DROP VIEW ['+isnull(@DestSchema, TABLE_SCHEMA)+'].[view_'+TABLE_NAME+'];

    EXEC(''CREATE VIEW ' + QUOTENAME( isnull(@DestSchema, TABLE_SCHEMA) ) + '.' + QUOTENAME( 'view_' + Table_Name )

    + ' AS' + CHAR( 10 )

    + 'SELECT ' + CHAR( 10 )

    + STUFF((

    SELECT ' ,' + QUOTENAME( t2.Column_Name ) + CHAR( 10 )

    FROM @TableInfo t2

    WHERE t2.Table_Name = t.Table_Name

    AND t2.Table_Schema = t.Table_Schema

    ORDER BY Ordinal_Position

    FOR XML PATH( '' )), 1, 5, ' ' )

    + ' FROM ' + QUOTENAME(@SourceDB) + '.' + QUOTENAME( Table_Schema ) + '.' + QUOTENAME( Table_Name )

    + ' SCHEMA_BINDING ' + CHAR( 10 ) + ''');'

    as Sql

    FROM @TableInfo t

    group by t.Table_Schema, t.Table_Name;

    declare @n int, @viewCount int;

    set @n = 1;

    select @viewCount = count(*) from @CreateViewStatements;

    DECLARE @sql As NVarchar(MAX);

    while @n <= @viewCount

    begin

    select @sql = v.Sql from @CreateViewStatements v where v.Id = @n;

    print @sql;

    exec(@sql);

    set @n = @n + 1;

    end

  • I'd strongly urge you to avoid the INFORMATION_SCHEMA views. Instead, stick to the sys. views.

    Not only are I_S views deficient in newer columns vs. the system tables, they are slow and lock/deadlock prone. Ah, you say, but when you list the definition of those views, it doesn't look that way. I know, but we have legacy issues with this all the time. I also believe there were cases where the I_S views didn't list certain objects, but it's been a while on that one so it may be resolved by now.

    SQL DBA,SQL Server MVP(07, 08, 09) "It's a dog-eat-dog world, and I'm wearing Milk-Bone underwear." "Norm", on "Cheers". Also from "Cheers", from "Carla": "You need to know 3 things about Tortelli men: Tortelli men draw women like flies; Tortelli men treat women like flies; Tortelli men's brains are in their flies".

  • One alternative to avoid INFORMATION_SCHEMA and use sys.tables and sys.columns instead , but for generation of views inside the same database than tables but using different schemas  (that must exists).
    The store procedure use DROP VIEW IF EXISTS synthax (so for 2016+ only or you will need to change this part)

    CREATE Procedure [usp_generate_view_script] (@SourceSchema VARCHAR(100) ,@TargetSchema VARCHAR(100))
    AS

        BEGIN
        DECLARE @Columns VARCHAR(MAX) = '';
        DECLARE @Schema VARCHAR(100);
        DECLARE @Table VARCHAR(200);
        DECLARE @TableID INT;
        DECLARE @SQLSELECT NVARCHAR(MAX)
        DECLARE @SQLCommand NVARCHAR(MAX);

        --CURSOR
        DECLARE curTables CURSOR LOCAL
        FOR SELECT SCHEMA_NAME(schema_id) TABLE_SCHEMA, name TABLE_NAME, object_id
        FROM sys.tables
        where TYPE='U' and SCHEMA_NAME(schema_id)=@SourceSchema
        FOR READ ONLY;

        OPEN curTables;
        FETCH NEXT FROM curTables INTO @Schema,@Table,@TableID;

        WHILE (@@fetch_status <> -1)
        BEGIN

        IF (@@fetch_status <> -2)
        BEGIN
        SET @Columns='';

        SELECT
         @Columns = @Columns + COALESCE('[' + NAME + ']' + ',', '')
        FROM
         sys.columns
        WHERE
         object_id = @TableID
        ORDER BY
         column_id;

        SET @Columns = SUBSTRING(@Columns, 1, LEN(@Columns) - 1);

        SET @SQLSELECT = N'SELECT ' + @Columns + ' FROM [' + @Schema + '].[' + @Table +'];';

        --PRINT @Sql

        PRINT @SQLCommand;

        BEGIN TRY

         SET @SQLCommand = N'DROP VIEW IF EXISTS [' + @TargetSchema + '].[' + @Table + '];';
         EXEC sp_executesql @statement=@SQLCommand ;
         SET @SQLCommand = N'CREATE VIEW [' + @TargetSchema + '].[' + @Table + '] AS ' + @SQLSELECT;
         PRINT @SQLCommand;
         EXEC sp_executesql @statement=@SQLCommand ;

        END TRY
        BEGIN CATCH

            DECLARE @ErrorMessage NVARCHAR(4000);
            DECLARE @ErrorSeverity INT;
            DECLARE @ErrorState INT;

            SELECT 
                @ErrorMessage = ERROR_MESSAGE(),
                @ErrorSeverity = ERROR_SEVERITY(),
                @ErrorState = ERROR_STATE();

            RAISERROR (@ErrorMessage, -- Message text.
                     @ErrorSeverity, -- Severity.
                     @ErrorState -- State.
                     );     

        END CATCH

        END

        FETCH NEXT FROM curTables INTO @Schema,@Table,@TableID

        END

        DEALLOCATE curTables

    END

  • You ABSOLUTELY do not want to use SELECT *.

    That alone with cause all kinds of performance problems.  There are multiple articles explaining why.  You will have to generate the list of columns to select.

    Theoretically one could generate the CREATE TABLE scripts for all the tables you want, then change the CREATE TABLE to CREATE VIEW, and then use Notepad++ to create editing macros to remove the parts you don't want, such as  data types, primary key defs, and other add-on scripts such as for constraints.

    Or you could use Python or Powershell to clean up the scripts.

  • Getting back to the original replies on this thread, which more than adequately expressed the right idea which I'll strongly second, this exercise is a total waste of time and will not prevent any slowdowns for reporting.  Whether you use SELECT * or not, it may actually cause several performance and resource usage problems but, even if it doesn't, this drill is a total waste of time and resources for the original problem.

    --Jeff Moden


    RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
    First step towards the paradigm shift of writing Set Based code:
    ________Stop thinking about what you want to do to a ROW... think, instead, of what you want to do to a COLUMN.

    Change is inevitable... Change for the better is not.


    Helpful Links:
    How to post code problems
    How to Post Performance Problems
    Create a Tally Function (fnTally)

  • @Flashx, You should ALWAYS use QUOTENAME() instead of hard-coding opening and closing brackets[].  One of the methods used in SQL injection is adding spurious closing brackets.

    Drew

    J. Drew Allen
    Business Intelligence Analyst
    Philadelphia, PA

  • SELECT 'CREATE VIEW ' + tb.TABLE_SCHEMA + '.vw' + tb.TABLE_NAME + ' AS SELECT '+

    (SELECT STUFF((SELECT ',t.' + vp.Column_Name FROM information_schema.columns vp

    WHERE vp.Table_Name=tb.TABLE_NAME and vp.Table_Schema=tb.TABLE_SCHEMA FOR XML PATH('')), 1, 1, ''))

    +' FROM '

    + tb.TABLE_SCHEMA +'.'+ tb.TABLE_NAME + ' t with(nolock)'

    FROM INFORMATION_SCHEMA.TABLES tb

    where TABLE_TYPE='BASE TABLE'

  • Links are not accessible 🙂 I'm quite interested in reading it.

Viewing 15 posts - 31 through 45 (of 52 total)

You must be logged in to reply to this topic. Login to reply