Enjoy!

You probably well aware that keeping statistics up-to-date is essential for SQL Server performance. You may also well know that if statistics is missing  on the columns of the table SQL Server creates _WA_Sys* stats for you automatically to have something instead of nothing to help optimizer decide which plan to chose. You shall not be relying on this in practice and always strive to keep your stats in a good shape. Script below will help you to convert numerous _WA_Sys* stats on your server into more meaningfully named single column stats. 

SELECT

N'

DROP STATISTICS [' + schema_name(t.schema_id)+N'].['+ t.name + N'].[' + s.name + N']

GO

CREATE STATISTICS [' + c.name + N'] ON [' + schema_name(t.schema_id)+N'].['+t.name + N'] ([' + c.name + N']) WITH FULLSCAN

GO

'

FROM sys.stats s (NOLOCK)

INNER JOIN sys.stats_columns (NOLOCK)sic ON sic.stats_id=s.stats_id AND sic.object_id=s.object_id

INNER JOIN sys.tables t (NOLOCK) on t.object_id = s.object_id

INNER JOIN sys.columns c (NOLOCK) on c.column_id = sic.column_id and c.object_id = sic.object_id

WHERE s.name like '%_WA_Sys%'

ORDER BY t.name

GO

Couple of days ago I had to move large table (~43 mln rows) from production environment to a dev environment.

Below is a table structure as follows:

CREATE TABLE [mmm].[Music Band](

[BandID] [int] IDENTITY(1,1) NOT NULL,

[BandTypeID] [tinyint] NOT NULL,

[BandName] [nvarchar](255) NOT NULL,

[IsAvailable] [bit] NOT NULL,

[IsRemoved] [tinyint] NOT NULL,

[CreatedDataSourceTypeID] [smallint] NOT NULL,

[CreatedDataSourceID] [int] NOT NULL,

[UpdatedDataSourceTypeID] [smallint] NOT NULL,

[UpdatedDataSourceID] [int] NOT NULL,

[DateCreated] [datetime] NOT NULL DEFAULT (getdate()),

[DateModified] [datetime] NOT NULL DEFAULT (getdate()),

[LastUpdated] [timestamp] NOT NULL

)

My initial approach was to use SSIS package to export data to a text file, then ship text file via ftp to a different server and then load text file via SSIS package to a table on a new server.

This approach failed miserably as it took very long time to debug package failures due to load issues from text file. Firstly, [BandName] can have different text data in it, including special characters like ],[,| %, $ etc since it represents name of the music band. So, no matter which delimiters you use in text file since there 43 mln rows in original table, likelihood of failure is very high. There are other possible reasons for failures, but I will omit them here.

So, what I did instead was to create a separate database, move table definition and all its data to this new database via SSIS (simple package with OLEDB Source and destination in data flow task), backup database, then ftp backup to the destination server, restore database backup on it and then move data to destination database again via SSIS (simple package with OLEDB Source and destination in data flow task), drop intermediate database and voila, done.

I hope you find this simple trick useful.

Cheers,

-Yuriy

Simple query below uses serverproperty function to tell you system level info about SQL Server.

I build it just recently for a small pet project of mine and decided to post it here just in case someone finds it useful too :)  

select 

serverproperty('ProductVersion') as [SQL server build number],

serverproperty('Edition') as [SQL server SKU],

serverproperty('ProductLevel') as [SQL server service pack level],

serverproperty('IsClustered') as [Is SQL server clustered],

serverproperty('MachineName') as [Windows server name],

serverproperty('ServerName') as [Full SQL server instance name],

serverproperty('InstanceName') as [SQL server instance name],

serverproperty('ProcessID') as [SQL server Windows process ID],

serverproperty('Collation') as [SQL server instance collation] 

Today I am going to chat briefly about the question a co-worker asked me about.

The problem was with embedding one XML snippet generated by using FOR XML option of the query into another XML also generated using FOR XML option.

Here is cleaned up version of the query :

declare @t table(c1 int, c2 int, ts binary(8))

insert into @t(c1,c2)

select 1, 2 

union all

select 3, 4 

union all

select 5, 6

declare @message xml

select @message =

(  select      

123 as '@attribute1',

                345 as '@attribute2',

456 as '@timestamp',

                getdate() as '@currentDate',

                     (select c1 as '@c1',

                             c2 as '@c2'

                  from @t

                       for xml path('item')

                      ) as 'items'

           for xml path ('items_list'), binary base64

      )

select @message 

If you execute this query directly against your instance of SQL Server you will get the following result:

<items_list attribute1="123" attribute2="345" timestamp="456" currentDate="2010-01-29T19:43:36.117">

  <items>&lt;item c1="1" c2="2"/&gt;&lt;item c1="3" c2="4"/&gt;&lt;item c1="5" c2="6"/&gt;</items>

</items_list> 

Please notice &lt; and &gt; in the output, since '<' and '>' have specific meaning in XML as tags. This is surely not what my coworker expected and intended.

The reason for this is that the result of the inner FOR XML query is returned as nvarchar(max) by default. The straightforward way around it is to cast the result to xml type using CAST as follows:

declare @t table(c1 int, c2 int, ts binary(8))

insert into @t(c1,c2)

select 1, 2 

union all

select 3, 4 

union all

select 5, 6

declare @message xml

select @message =

(  select      

123 as '@attribute1',

                345 as '@attribute2',

 456 as '@timestamp',

                getdate() as '@currentDate',

                     cast (

(select c1 as '@c1',

                             c2 as '@c2'

                  from @t

                       for xml path('item') 

                      )

as xml ) as 'items'      

           for xml path ('items_list'), binary base64

      )

select @message 

The output of this is what was intended:

<items_list attribute1="123" attribute2="345" timestamp="456" currentDate="2010-01-29T19:51:06.940">

  <items>

    <item c1="1" c2="2" />

    <item c1="3" c2="4" />

    <item c1="5" c2="6" />

  </items>

</items_list>

 However, in SQL Server FOR XML clause allows for TYPE option to achieve exactly this so there is no need to cast the result to XML type:

 declare @t table(c1 int, c2 int, ts binary(8))

insert into @t(c1,c2)

select 1, 2 

union all

select 3, 4 

union all

select 5, 6

declare @message xml

select @message =

(  select      

123 as '@attribute1',

                345 as '@attribute2',

   456 as '@timestamp',

                getdate() as '@currentDate',

                     (select c1 as '@c1',

                             c2 as '@c2'

                  from @t

                       for xml path('item'), TYPE 

                      ) as 'items'

           for xml path ('items_list'), binary base64

      )

select @message

with much desired output:

 <items_list attribute1="123" attribute2="345" timestamp="456" currentDate="2010-01-29T19:55:37.973">

  <items>

    <item c1="1" c2="2" />

    <item c1="3" c2="4" />

    <item c1="5" c2="6" />

  </items>

</items_list>

Hope description of this small 'gotcha' was useful.

Until next time.

-Yuriy  

select qp.query_plan,

SUBSTRING(st.TEXT, (qs.statement_start_offset/2)+1 ,

(( CASE qs.statement_end_offset

WHEN -1 THEN datalength(st.TEXT)

ELSE qs.statement_end_offset

END

- qs.statement_start_offset)/2) + 1) AS txt

from sys.dm_exec_query_stats qs

cross apply sys.dm_exec_sql_text(sql_handle) st

cross apply sys.dm_exec_query_plan(plan_handle) qp

Then one can modify above statment with XQuery to query for Missing Index information using above mentioned XSD schema for XML showplan in SQL Server 2008 to get what potential index is missing for the statement. Please note that not all statement have index missing, only some.   

Luckily, Brent Ozar and great site www.sqlserverpedia.com already have full query ready for this which I copied down below for your convenience.

SELECT qp.query_plan
, total_worker_time/execution_count AS AvgCPU
, total_elapsed_time/execution_count AS AvgDuration
, (total_logical_reads+total_physical_reads)/execution_count AS AvgReads
, execution_count
, SUBSTRING(st.TEXT, (qs.statement_start_offset/2)+1 , ((CASE qs.statement_end_offset WHEN -1 THEN datalength(st.TEXT) ELSE qs.statement_end_offset END - qs.statement_start_offset)/2) + 1) AS txt
, qp.query_plan.value('declare default element namespace "http://schemas.microsoft.com/sqlserver/2004/07/showplan"; (/ShowPlanXML/BatchSequence/Batch/Statements/StmtSimple/QueryPlan/MissingIndexes/MissingIndexGroup/@Impact)[1]' , 'decimal(18,4)') * execution_count AS TotalImpact
, qp.query_plan.value('declare default element namespace "http://schemas.microsoft.com/sqlserver/2004/07/showplan"; (/ShowPlanXML/BatchSequence/Batch/Statements/StmtSimple/QueryPlan/MissingIndexes/MissingIndexGroup/MissingIndex/@Database)[1]' , 'varchar(100)') AS [DATABASE]
, qp.query_plan.value('declare default element namespace "http://schemas.microsoft.com/sqlserver/2004/07/showplan"; (/ShowPlanXML/BatchSequence/Batch/Statements/StmtSimple/QueryPlan/MissingIndexes/MissingIndexGroup/MissingIndex/@Table)[1]' , 'varchar(100)') AS [TABLE]
FROM sys.dm_exec_query_stats qs
cross apply sys.dm_exec_sql_text(sql_handle) st
cross apply sys.dm_exec_query_plan(plan_handle) qp
WHERE qp.query_plan.exist('declare default element namespace "http://schemas.microsoft.com/sqlserver/2004/07/showplan";/ShowPlanXML/BatchSequence/Batch/Statements/StmtSimple/QueryPlan/MissingIndexes/MissingIndexGroup/MissingIndex[@Database!="m"]') = 1
ORDER BY TotalImpact DESC

So, now you know what indexes are missing and also have an idea what SQL statements benefit from that index.

Happy querying!

 -Yuriy 

 Set oWorkBook = oExcel.Workbooks.Open("<path to your excel file>")

Now you saved yourself a whole lot time and can start working on more interesting things!

Cheers, -Yuriy

Script below

Function Main()
Dim oExcel
Dim oWorkBook
Dim oWorksheet
Dim oPivotTable
Dim IStartedExcel
 
On Error Resume Next
Set oExcel = GetObject("Excel.Application")       'will attach to an existing instance if it is there.
oExcel.Quit
 
'Instantiate excel object; disable alert messages
Set  oExcel = CreateObject("Excel.Application")
oExcel.DisplayAlerts = false
 
'when debugging set visible = true
oExcel.visible = true

'open workbook we wish to refresh
Set oWorkBook = oExcel.Workbooks.Open("<path to your excel file>")
oWorkbook.EnableConnections
oWorkbook.RefreshAll
oExcel.CalculateUntilAsyncQueriesDone
oWorkbook.Save
oWorkBook.Close
 
oExcel.Quit
Set oExcel = Nothing
 
'Main = DTSTaskExecResult_Success

End Function