FileGroup

Hi,
There is a table which is about 50GB is size.
I am thinking of placing the database of this table into a separate .ndf on a separate drive on the server.
Is it possible to place this particular table into a searate drive to increase the performance even more?
Basically my question is: is it possible to place tables into separate .ndf files?
ThanksYes, Microsoft did a great whitepaper about exactly this thought a few years ago. I'm not sure if they made it public or if it was "restricted publication" due to containing customer specific details.

The general gist of the idea is to create specific filegroups for specific devices, then put the SQL objects on those filegroups.

One area that often helps more than moving specific tables to a given drive is to segregate data pages (either the heap or the clustered index) onto one set of spindles (aka drive) and the indicies onto a different filegroup. This allows parallel IO to occur for even single table changes (assuming you have enough CPUs and IO channels)!

-PatP|||I think I am now confused.
Could you please elaborate further and do you know of a link regarding this issue?
Thanks|||We can create two filegroups or more while creating database. The following SQL Statements maybe help you:

CREATE DATABASE database_name
[ ON
[ PRIMARY ] [ <filespec> [ ,...n ]
[ , <filegroup> [ ,...n ] ]
[ LOG ON { <filespec> [ ,...n ] } ]
]
[ COLLATE collation_name ]
[ WITH <external_access_option> ]
]

Here in the block <filespec>,you have to give the file path,then you can locate the files into different disk.

Then referencing the following SQL statements:

CREATE TABLE
[ database_name . [ schema_name ] . | schema_name . ] table_name
( { <column_definition> | <computed_column_definition> }
[ <table_constraint> ] [ ,...n ] )
[ ON { partition_scheme_name ( partition_column_name ) | filegroup
| "default" } ]
[ { TEXTIMAGE_ON { filegroup | "default" } ]
[ ; ]

So you store datum in different disk.

Good luck!|||Yes, this is what I was after.
This seems to be a very powerfull stuff.
Thank you