Newsequentialid (Histrory/Benefits and Implementation)

re: Newsequentialid (Histrory/Benefits and Implementation)

mobilebilly — Thu, 30 Nov 2006 06:49:08 GMT

We are considering to use the OS function UuidCreateSequential to generate the key at the client side.

As you said the implementation of the newsequentialid in Sql Server 2005 scramble the result from UuidCreateSequential, is there any issue with that approach?

Is there any detail information about the scrambling?

A Class that Creates Sequential GUIDs

Scott Bellware [MVP] — Wed, 27 Dec 2006 05:26:55 GMT

My current project uses GUID primary keys in a SQL Server 2005 database. The primary keys are created

re: Newsequentialid (Histrory/Benefits and Implementation)

BSovers — Mon, 29 Jan 2007 23:19:59 GMT

It is frustrating implementing NewSequentialID() as a default value within a table (though that is the ONLY place it CAN be used). Every time I try it, SQL SMS keeps giving an "Error validating the default for column..." message each time I attempt to apply it.

re: Newsequentialid (Histrory/Benefits and Implementation)

MorningZ — Tue, 13 Feb 2007 18:34:13 GMT

Any luck ever getting around the validating error?

NewSequentialId() in C#

Hugo Ribeiro — Sun, 22 Jul 2007 18:27:36 GMT

A colleague pointed me to this code that implements NewSequentialId() in C#: http://codebetter.com/blogs/scott.bellware/archive/2006/12/27/156671.aspx

re: Newsequentialid (Histrory/Benefits and Implementation)

zootie — Fri, 27 Jul 2007 19:14:24 GMT

Scott Bellware's c# implementation over on codebetter (link above) only calls the underlying OS function. However, SQL 2005's implementation does some byte re-ordering after calling the same function, and if you want the sequential GUIDs to be "compatible", you'd have to do the same operations on the client side. You can see the reordering in the source code of yaflaGUID (http://www.yafla.com/dforbes/stories/2005/10/17/yaflaguid.html). You'd have to translate the assembly code to c#. In essence you call UuidCreateSequential (as Scott does), then invoke ToByteArray, swap bytes 0 & 3, 1 & 2, 4 & 5 and 6 & 7, and create a new GUID from the byte array.

re: Newsequentialid (Histrory/Benefits and Implementation)

KCCE — Thu, 30 Aug 2007 00:46:38 GMT

I posted a C# implementation that handles the byte reordering as a comment to Scott Bellware's Blog post.

Go to Scott's blog entry (link above) and look in the comments for the c# snippet. This will produce a guid that is equivalent to SQL's newSequentialID(). I did place the external definition to UuidCreateSequential() in a SafeNativeMethods class that is not included in the snippet.

cheers.

Kendall

re: Newsequentialid (Histrory/Benefits and Implementation)

ShaneAKenyon — Tue, 16 Oct 2007 23:59:52 GMT

The validation error seems to be a SQL Studio bug. You can still save the table (after accepting the error _again_), using the function as the default value. Thanks Scott and Kendall for your code - I will be working on a NHibernate generator for this and hopefully post it to my blog soon.

re: Newsequentialid (Histrory/Benefits and Implementation)

ha-man — Fri, 08 Feb 2008 12:52:21 GMT

Just an observation on the re-ordering of bytes in the implementation of newsequentialid. I understand fully why this must be performed to adhere to the ordering algorithm in SQL Server.

But as I understand it this changes the algorithm version for generating the guid from 1 (which is what UuidCreateSequential uses), to whatever is currently the 2nd highest 4 bits of the time (currently D on my system).

Am I correct in assuming that this is currently not a problem in practice because (i.e. it will not cause clashes with other algorithms because):

1) There is no standardized algorithm which has been assigned D, E or F as version number

2) The 2nd highest 4 bits of the time will not wrap around to a lower value any time soon.

Unraveling the mysteries of NewSequentialID

Technical Thoughts of Jorriss — Thu, 24 Apr 2008 18:56:07 GMT

A few months ago we started development on a new system. From the ground up we redesigned everything

Unraveling the mysteries of NewSequentialID

Technical Thoughts of Jorriss — Sun, 26 Oct 2008 18:06:36 GMT

Unraveling the mysteries of NewSequentialID

Sequential GUID i SQL Server og i .NET

creave.dk — Wed, 13 May 2009 13:15:01 GMT

Sequential GUID i SQL Server og i .NET

GUID Ordering in SQL Server

Martin Bell UK SQL Server MVP — Mon, 25 May 2009 16:38:18 GMT

When creating client-side guids to be stored in SQL Server things are not as simple as you expect...

SQL Programmability amp API Development Team Blog Newsequentialid | Weak Bladder

SQL Programmability amp API Development Team Blog Newsequentialid | Weak Bladder — Mon, 08 Jun 2009 04:22:29 GMT

PingBack from http://weakbladder.info/story.php?id=387

SQL Programmability amp API Development Team Blog Newsequentialid | Toe Nail Fungus

Wed, 10 Jun 2009 05:15:25 GMT

PingBack from http://toenailfungusite.info/story.php?id=4164