Partitioning on a Table - Few Questions, Confusions

Question

Hello All,
   I've a table with around 300 Million Records. This Table has one Key Column (Auto Incremented by 1 ) and a Unique ID Field.
Primary Key (Clustered Primary Key) was created on Key Column of this Table.
Now If I want to partition this table based on UniqueID Field, SQL Server is not allowing me to do it! It throws the error something like "Partitioned Column should present in Primary Key (Clustered Index)"
Is it mandatory that the partitioned column should definitely be present in the clustered index or primary key of the table if exists.
Could someone tell me what are the prerequisites to partition a table that already have a primary key on FieldA and Unique Non Clustered Index on FieldB and I'd like to partition the table based on FieldC?
Thanks in Advance...
Unknown

Answer 1

Hi Roger,
I am a little confused, why do you want to partition a table based on a UniqueID? Partioning is more of a logical thing to do like dividing a table yearly, for ex the table can be divided on a year by year basis which would make more sense as the older data
can be moved over to slow storage if it's used as often as the new one based on the environment.
A UniqueID would just be a random number, I am not sure how are you planning to create a partition scheme on this and would that be effective.
Just to answer your question, it's not mandatory to have a PK column included in the partitioning.
For Ex, one of the scripts that I wrote
--CREATE PARTITION FUNCTION PF_test (datetime)
AS RANGE LEFT FOR VALUES
 ('2010-12-31 23:59:59.999',
 '2011-12-31 23:59:59.999',
 '2012-12-31 23:59:59.999',
 '2013-12-31 23:59:59.999',
 '2014-12-31 23:59:59.999',
 '2015-12-31 23:59:59.999')
GO
I would suggest that you watch the below video to have a better understanding:-
http://technet.microsoft.com/en-US/sqlserver/gg545008.aspx
Please mark the answer as helpful if i have answered your query. Thanks and Regards, Kartar Rana