June 01, 2002 07:01 PM

Just in Time

Rating: (0)
SQL Server Magazine
InstantDoc ID #25174
H.H. Munro, writing under the pseudonym of Saki in late 19th-century Britain, opined, "A little inaccuracy sometimes saves tons of explanation." Unfortunately, a discussion about datetime data's inaccuracy is necessary, even though the data is only a little inaccurate. Although a datetime data type's time can be precise to milliseconds, SQL Server can't represent every millisecond. In fact, within the millisecond value of the datetime data type, SQL Server rounds to within 0.00333 seconds, a pe...

...This article is for paid Professional Members only.

Already a Professional Member? Please log in now:

NOT A PROFESSIONAL MEMBER? YOU CHOOSE:

Professional Membership

Monthly

Annual

VIP Membership

Monthly

Annual

Add a Comment

Again, this information is available on BOL. Kimberly, please, stop writing.

Marcos3/18/2009 2:14:37 PM


I know this is an older article, but I wanted to know what anyone thought about using a datetime column as a clustered index.

I realize the overhead of using a non-unique field, but the only other key fields are all GUIDs (using NEWID() because this was originally designed for SQL 2000).

The GUIDs are have unique indexes on them and are set up as Foreign Keys in other tables.

The apps using this design is about to be rewritten so I've been asked to help redesign the schema. My instinct says to add an Identity column and make that Primary Key and the Clustered Index, but I'm at odds about whether I should stick with the current design.

Phil4/12/2007 9:07:34 AM


You must log on before posting a comment.

Are you a new visitor? Register Here
GOOGLE LINKS
SPONSORED LINKS
FEATURED LINKS