April 2008 - Posts

SQL Server 2008 UK Usergroup London Launch event

As part of the SQL Server Launch wave we are holding a usergroup meeting to celebrate the launch (not RTM).

We're going to run this as an open session so you'll be able to ask us to cover the areas you want us to cover.

We have loads of giveaways, including Vista, training vouchers, Technet subscriptions and SQL Server licenses.
To register for the event go to the UK SQL Usergroup site http://www.sqlserverfaq.com/?eid=114

Registration is at 5.30, evening will commence at 6pm and finish 9pm.

6pm - 6.30pm
Round Table Discussion
Update on what's been going on and is going on in the SQL Server space.
Bring your SQL problems and ask the audience, bounce ideas - anything related to SQL Server.

The topics to cover in next parts could be any of the following, Simon and Jasper will hope to answer and demo the features you want to know about.

TSQL improvements, new data types, changes to CLR, spatial data, hierarchies, service broker, changes to tools, SSIS improvements, Integrated Full Text, Sparse Columns, Filtered Indexes, XQuery changes, Compression, Change Data Capture, Change Tracking, Intellisense, Table Valued Parameters, Script Task in SSIS, Performance Data Collector, Reporting services

6.30pm – 7.30pm 
Part 1

7.30 – 7.50pm 
Break - Beer and Pizza

7.50pm - 9.00pm
Part 2

Speaker Bios
Simon Sabin  http://sqlblogcasts.com/blogs/simons Jasper Smith http://sqlblogcasts.com/blogs/sqldbatips
Simon is Director of his own data consultancy company Onarc Consulting that specialises in data related systems. He is also a Developer Skills Partner with SQLSkills. He has a particular expertise in the world of search, distributed architectures, business intelligence and application development. He has worked with SQL Server since 1998 and has always focused on high performance reliable systems.
Simon was awarded as an MVP in 2006. He founded the first free SQL Server conference in the UK SQLBits in 2007 along with other MVPs in the UK. He is a regular speaker at SQL Server events as well as writing his blog and writing for SQLServerCentral and Simple-Talk.com.
He has most recently been involved in developing the SQL Server 2008 training for Microsoft.
Jasper has been working with SQL Server for around 6 years and is currently a member of the Database Services team at Nationwide Building Society in the UK. He also writes articles and code for his website : http://www.sqldbatips.com
Posted by simonsabin | with no comments
Filed under: , , ,

Don't forget the London SQL Usergroup is meeting this Thurdsay (17th April).

The agenda is as follows,
6pm - 6.30pm
Round Table Discussion
Update on what's been going on and is going on in the SQL Server space.
Bring your SQL problems and ask the audience, bounce ideas - anything related to SQL Server.

6.30pm – 7.30pm  Will you still want me, when I'm 64? - Christian Bolton (COEO)
An in-depth look at the move toward 64-bit computing with SQL Server covering where and why it will help, what the caveats are and the most up to date best practices gathered from global SQL Server deployments.

7.30 – 7.50pm Break

7.50pm - 9.00pm
The cursors of SQLStein - Eric Alsop (IM Group)

Nothing divides the SQL community like a cursor, so why are they still around? In this session we are going to look at how to use cursors and when they are most applicable.

To register go here http://www.sqlserverfaq.com/?eid=110

Posted by simonsabin | 1 comment(s)

If you are using CTP6 of SQL Server 2008 and want to use Integrated Full Text and find it is not working. You might get timeouts or other general errors. It could be due to an account that iFTS uses.

iFTS does some stuff out of process for security and reliability in CTP6 the run this process under a local user account FDH$<instance name> by default.

However this accounts password may expire or even may be set to change on next logon (depending on domain settings etc).

If you find iFTS is not working then go into Local Users and Groups. Select the FDH$ account and check the settings. If you need to change the password there is a sp sp_fulltext_resetfdhostaccount which you can use to sync the password in SQL once you have changed it in the User manager.

 THIS ONLY APPLIES TO CTP 6 AND PRE CTP6. I've been told the behaviour is changing for future CTPs.

If you've got a complex data set to produce how should you do it. Adam Machanic has posted about an Anti Pattern (bad practice) for such a situation. http://sqlblog.com/blogs/adam_machanic/archive/2007/12/15/anti-patterns-and-malpractices-volume-1-tumbling-data.aspx

He talks baout the approach of building temp tables and then using the them to produce more temp tables and finally the result. I agree in many situations this is a bad practice however as with database normalisation there is a trade off which you need to decide upon.

I personally find that building individual statements can be good for a few reasons,

  • Supporting the code.
    Derived tables, pivots, case statements and sub queries are advanced SQL and so using lots of them means the person supporting the code has to be able to understand it.

  • Debuging code
    If you have an issue with the output having the code in one bunch generally means you have to disect the query to be able to debug it. Having it already broken down eases this issue and allows for assertions to be made in the code to ensure that the data is behaving as expected.

  • Performance
    Putting everything in one query will greatly increase the options for the optimiser. I have seen this result in very bad plans, especially with parallism involved.

  • Manageability
    Doing things in one means you have a single black box, it starts and finishes. Its very difficult to see what going on. Breaking it up into smaller chunks provides better options for seeing and managing progress.

  • Memory usage
    Doing it all in one go can result in you needing a large amount memory for the whole query. Splitting it up reduces the single memory requirements.

Now I have to say that these comments are generally when building large complex data sets for reports or exporting data. In addition there is also the ETL scenario. In that case you have a decision on the granularity of your processing whether you want row by row or set based. The last generally offers best performance but less control over errors.

So whilst I agree with Adam there is an element of it depends on the situation.

Posted by simonsabin | with no comments

I've been hitting F9 in outlook all day waiting for the email to tell me if I had been reawarded. It finally came just before the letters were worn off the F9 key :)

If you want to know more about the MVP programme you can read more at the MVP site http://mvp.support.microsoft.com/Default.aspx.

Posted by simonsabin | 1 comment(s)

Sorry about the scripting and layout problems over the past week or so. Since we upgraded to community server 2007 some of the custom stuff I have broke.

It should all be now fixed.

If you find any more problems let me know.

Posted by simonsabin | 1 comment(s)