SQL Server

drop and recreate sql server indexes

Alex Bransky — Wed, 04 Jan 2012 21:22:08 GMT

I have a real-time reporting server with a database which is basically a copy of my production database, but it's populated using transactional replication and replication only copies the primary keys over, i.e. none of the secondary indexes are included. So what I do is every time I need to recreate replication--which is every deployment with schema changes--I script out DROPs and CREATEs for all the reporting indexes I've created and then run the script after replication has been recreated. The script I use is a modified conglomeration of scripts I've found on the Internet, but you may find it useful. Note that it excludes primary keys and some other things.

PRINT '------------------------- DROP INDEXES     ----------------------------------------------------'
DECLARE @ownername SYSNAME
DECLARE @tablename SYSNAME
DECLARE @indexname SYSNAME
DECLARE @sql NVARCHAR(4000)
DECLARE dropindexes CURSOR FOR
    SELECT indexes.name, objects.name, schemas.name
    FROM sys.indexes
    JOIN sys.objects ON indexes.OBJECT_ID = objects.OBJECT_ID
    JOIN sys.schemas ON objects.schema_id = schemas.schema_id
    WHERE indexes.index_id > 0
    AND indexes.index_id < 255
    AND objects.is_ms_shipped = 0
    AND NOT EXISTS (SELECT 1 FROM sys.objects WHERE objects.name = indexes.name)
    ORDER BY indexes.name

--SELECT * FROM sys.stats
OPEN dropindexes
FETCH NEXT FROM dropindexes INTO @indexname, @tablename, @ownername
WHILE @@fetch_status = 0
BEGIN
SET @sql = N'IF EXISTS (SELECT 1 FROM sys.indexes i WHERE i.name = ''' + @indexname + ''') DROP INDEX '+QUOTENAME(@ownername)+'.'+QUOTENAME(@tablename)+'.'+QUOTENAME(@indexname)
PRINT @sql
FETCH NEXT FROM dropindexes INTO @indexname, @tablename, @ownername
END
CLOSE dropindexes
DEALLOCATE dropindexes

PRINT CHAR(10) + CHAR(13) + '------------------------- CREATE INDEXES     ----------------------------------------------------'

declare
@object_id          int,
@index_id           tinyint,
@schema_name        sysname,
@table_name         sysname,
@index_name         sysname,
@type               tinyint,
@uniqueness         bit,
@indexed_column     sysname,
@included_column    sysname,
@indexed_columns    varchar(max),
@included_columns   varchar(max),
@has_included_cols bit,
@is_descending_key bit,
@stmt               varchar(max),
@crlf               char(2)

set @crlf = char(13) + char(10)

declare indexes cursor
for
    select
    schema_name     = s.name,
    table_name      = t.name,
    index_id        = i.index_id,
    index_name      = i.name,
    type            = i.type,
    uniqueness      = i.is_unique
from
         sys.schemas            s
    join sys.tables             t   on s.schema_id = t.schema_id
    join sys.indexes            i   on t.object_id = i.object_id
where
    i.type > 0 -- none -heap
    AND I.index_id > 0 -- ignore PKs
    AND I.index_id < 255
    AND i.type = 2
    AND INDEXPROPERTY (I.object_id,I.NAME,'ISCLUSTERED') =0
    AND left(I.object_id,3) not in ('sys', 'dt_')
    AND left(I.name,3) not in ('PK_', 'UQ_','nci','ucm','ci_','uk_')
order
    by i.name

open indexes

fetch
    indexes
into
    @schema_name,
    @table_name ,
    @index_id   ,
    @index_name ,
    @type       ,
    @uniqueness

while @@fetch_status<>(-1)
begin

    select @object_id = object_id(@schema_name + '.' + @table_name)
    set @indexed_columns = '('

    declare indexed_columns cursor
    for
        select
            c.name,
            ic.is_descending_key
        from
                sys.index_columns ic
            join sys.columns        c   on ic.column_id = c.column_id
                                       and ic.object_id = c.object_id
        where
            ic.object_id = @object_id
            and ic.index_id = @index_id
            and ic.is_included_column = 0
        order by
            ic.index_column_id

    open indexed_columns

    fetch indexed_columns
    into @indexed_column, @is_descending_key

    while @@fetch_status<>(-1)
    begin

        set @indexed_columns = @indexed_columns + @indexed_column +
                               case @is_descending_key when 1 then ' DESC ' else '' end + ', '

        fetch indexed_columns
        into @indexed_column, @is_descending_key

    end

    close indexed_columns
    deallocate indexed_columns

    set @indexed_columns = left(@indexed_columns, len(@indexed_columns)-1) + ')'

    if exists
           (select object_id
            from    sys.index_columns
            where   object_id = @object_id
                    and index_id = @index_id
                    and is_included_column = 1 )
    begin
        set @included_columns = 'INCLUDE ('

        declare included_columns cursor
        for
        select
            c.name,
            ic.is_descending_key
        from
                 sys.index_columns ic
            join sys.columns        c   on ic.column_id = c.column_id
                                       and ic.object_id = c.object_id
        where
            ic.object_id = @object_id
            and ic.index_id = @index_id
            and ic.is_included_column = 1
        order by
            ic.index_column_id

        open included_columns

        fetch included_columns
        into @included_column, @is_descending_key

        while @@fetch_status<>(-1)
        begin

            set @included_columns = @included_columns + @included_column +
                                    case @is_descending_key when 1 then ' DESC ' else '' end + ', '

            fetch included_columns
            into @included_column, @is_descending_key

        end

        close included_columns
        deallocate included_columns

        set @included_columns = left(@included_columns, len(@included_columns)-1) + ') '

    end

    set @stmt =
    'CREATE ' + case @uniqueness when 1 then 'UNIQUE ' else '       ' end + case @type when 1 then 'CLUSTERED ' else '' end +
    'INDEX ' + @index_name + ' ON ' + @schema_name + '.' + @table_name + @indexed_columns + ' ' + isnull(@included_columns,'') + @crlf

    set @included_columns = ''

    print @stmt

    fetch
        indexes
    into
        @schema_name,
        @table_name ,
        @index_id   ,
        @index_name ,
        @type       ,
        @uniqueness

end

close indexes
deallocate indexes

annoying SSRS bug fix for building

Alex Bransky — Thu, 30 Jun 2011 18:19:58 GMT

If you're like me you've been using SSRS 2008 for a long time with TFS and have constantly had to delete the bin folder in order for builds to work. This has irritated me to no end. Well today is my lucky day! There is a hotfix that actually fixes the bug! Go here: http://support.microsoft.com/kb/2438347

indexes and deadlocks

Alex Bransky — Fri, 24 Jun 2011 18:27:03 GMT

I recently discovered that creating indexes in one of my production databases was causing deadlocks. My problem was that I wasn't using ONLINE=ON when I was creating the index. Check out the BOL for more information.

removing an ssrs instance from a scale-out deployment

Alex Bransky — Thu, 03 Mar 2011 19:49:30 GMT

If you're like me you had at one time connected one of your Reporting Services instances to a report server database that was already in use by another instance. This allows the instance to show up in the Scale-out Deployment section of the Reporting Services Configuration Manager. My problem was that the server that got joined to the original server was no longer available as it had been repurposed, and when I clicked Remove Server to remove it from my scale-out it would fail because it couldn't contact the server. After searching for a solution for quite some time I decided to look around in the report server database tables, and voila! All I had to do was remove the old server from the Keys table. I can't guarantee there won't be any side effects to this method, but it worked like a charm for me.

migrating sharepoint databases

Alex Bransky — Wed, 16 Feb 2011 22:55:44 GMT

If you're wondering how to migrate your SharePoint databases to a new server, this Microsoft article is actually pretty useful, though still overly complex like most of their other articles.

http://technet.microsoft.com/en-us/library/cc512725.aspx

The one thing I would change is that they seem to recommend installing SQL Server Configuration Manager on web servers, when all that was needed in my case was to add an entry to the hosts file on the SharePoint web server that used the IP address of the new SQL Server with the name of the old SQL Server. This might not be appropriate in cases where the old server is not being decommissioned.

workspaces in sql mgmt studio

Alex Bransky — Wed, 01 Jul 2009 22:46:09 GMT

How is it that you still can't save workspaces in SQL Server Management Studio as of version 2008? It seems like such a simple thing to implement. Quite often I create a bunch of different queries during a short period that are specific to a task I'm working on and that won't be needed once the task is completed, and rather than having to save each one it would be nice if I could just save a workspace file that knows which queries were open.

attaching adventureworks2008 on full version of sql server

Alex Bransky — Wed, 10 Jun 2009 00:39:05 GMT

For anybody that bought the SQL Server 2008 Self-Paced Training book for 70-433, if you are running a full version of SQL Server 2008 rather than the express edition, you will have problems attaching the AdventureWorks2008 database (file activation error). The trick is to use a SQL command, like so:

USE [master]
GO
CREATE DATABASE [AdventureWorks2008] ON
( FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL10.SQL2008\MSSQL\DATA\AdventureWorks2008_Data.mdf' ),
( FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL10.SQL2008\MSSQL\DATA\AdventureWorks2008_Log.ldf' ),
FILEGROUP Documents CONTAINS FILESTREAM (NAME = Documents, FILENAME=N'C:\Program Files\Microsoft SQL Server\MSSQL10.SQL2008\MSSQL\DATA\Documents')
FOR ATTACH
GO

tinyint surprise

Alex Bransky — Wed, 11 Feb 2009 01:05:03 GMT

I just discovered something rather surprising. If you return a tinyint from a stored procedure as part of a dataset, and bind that dataset to a dataview, and then do something like this:

<asp:Label runat="server" id="statusLabel" Text='<%# GetStatusText( (int)DataBinder.Eval(Container, "DataItem.status")) %>'>
</asp:Label>

You will get an error about an invalid cast. If the stored procedure returns an int instead of tinyint, it works just fine. How can ASP.Net 3.5 not be able to cast from a tinyint (or whatever its C# equivalent is) to an int??

activerecord and executereader exceptions

Alex Bransky — Wed, 28 May 2008 20:46:45 GMT

About once a day I was getting the following error from an ASP.Net web page that uses ActiveRecord:

Exception Details: System.Data.SqlClient.SqlException: A transport-level error has occurred when sending the request to the server. (provider: TCP Provider, error: 0 - An existing connection was forcibly closed by the remote host.)

I would subsequently get this error upon refreshing the page:

Exception Details: System.InvalidOperationException: ExecuteReader requires an open and available Connection. The connection's current state is closed.

I finally figured out that the problem occurs when the database is restored. The connections in the IIS application pool do not get destroyed when the database is restored, so when ActiveRecord (nHibernate) tries to reuse the connections it fails. This doesn't happen to me when I'm not using ActiveRecord, so I'm not sure if this is a bug with ActiveRecord or a problem with the way I've set up the data access layer.

ssas: define reference and materialize

Alex Bransky — Thu, 10 Apr 2008 22:25:34 GMT

Apparently you have to watch out when you select Materialize in a referenced dimension in SSAS 2005. I just ran into a situation where data was getting duplicated for no good reason, and unchecking Materialize fixed it.

http://www.biblogs.com/2006/12/23/materialize-option-in-a-reference-dimension/