Re:Database development mistakes made by application developers
by binarymax on Monday December 13, 2010 @08:28AM
(#2000376)
Good Additions. I would also point out that if devs mess with indexes they definitely need to know the difference between clustered and nonclustered. I've seen atrocities due to a simple misunderstanding of this (guess what happens when you confuse the two and implement the opposite!)
[ Parent
| on HN ]
Re:Database development mistakes made by application developers
by tocomment on Monday December 13, 2010 @09:30AM
(#2000535)
Use non-clustered, right? clustered is just for massive operations with distributed databases and junk, right?
Re:Database development mistakes made by application developers
by binarymax on Monday December 13, 2010 @09:38AM
(#2000559)
Clustered is directly related to the physical order of the records. Non-clustered is a logical order that uses pointers to the physical location. Only one clustered is allowed per table.
Re:Database development mistakes made by application developers
by ams6110 on Monday December 13, 2010 @09:46AM
(#2000582)
The concept also does not exist in some databases. Oracle, IIRC, does not have the notion of a clustered index. SQL Server does. Another illustration of why you should not assume that all relational databases work the same way.
Re:Database development mistakes made by application developers
by binarymax on Monday December 13, 2010 @10:01AM
(#2000631)
DB2 supports clustering, and Oracle has a similar concept:
http://www.oracle-base.com/articles/8i/IndexOrganizedTables....
You are definitely right about each RDBMS having their own quirks though! Solid advice: know what you are working with and its nuances.