+972 52-548-6969
A developer perspective on SQL Server 2008
time to read 1 min | 174 words
I just read this about the SQL Server 2008 features, and I thought to provide my own commentary on the new features:
- Declarative Management Framework - not really interesting to me.
- Table Value Parameters - YEAH, that is more like it. There are too many cases when I need to use IN on a large data set and I run into the 2100 parameters limitations. This is going to make working with this so much easier.
- MERGE SQL Statement - I got tired of writing my own Update/Insert very early. This also means that the tools will support it as well, probably (SSIS), so that makes my life easier still.
- Change Data Capture - Hm, that looks interesting in the abstract, will probably want to check that out.
- Star Join Query Optimizations - not really interesting yet.
- There is also the EDM stuff that they are apparently putting there, which is also interesting, but not likely to be usable unless you are using the Entity Framework. The announcement about that was very vague.
Comments
Interesting... what if... our poco definitions were defined and managed by the database?
It would then be SQL Server's job to hydrate and persist objects, handle change managements, and automatically handle transactions, maybe bypass some of that marshalling between odbc and sql data reader?
My bet is that it would allow for tighter control of transactions, less I/O, and a higher control over your caching and loading considerations.
I wouldnt know where to begin on this one, but it seems as if the school of DDD could see some value in this idea.
Just wanted to bounce that off someone who could offer insight
-p
What about Spatial Data, what is your opinion on that?
http://www.codeplex.com/NHibernateSpatial
SQL Server 2008 has 2 new spatial data types built in
Nhibernate uses MsSqlSpatial which are Spatial Extensions for SQL Server 2005.
@Denis,
When it comes the time to use that, it would be easy to move it to the native 2008 support.
What I'd very much like to see in SQL Server is named sequences similiar to Oracle's. How difficult can that be? And it would save so much grief IDENTITY columns cause!
I wonder if there's some features request list we could vote on...
ET,
Sure there is a feature list, you just need to be CEO of Fortune 5 to do it, I think :-)
Comment preview