I found the data partitioning article interesting and tried to implement for my appliaction. The impletmentation discussed in this article is done.

My doubt is regarding the deletion part, in the site “http://msdn.microsoft.com/en-us/library/ms345146(SQL.90).aspx “, after creating a staging table the older data is switched and later the stagign table is dropped…
Can you suggest any alternative mthod available to drop partion with lesaa complexity.

Thanks in advance
Rupa

I have gone through your tutorial and havinf sme question. I have one table where it’s aroung 500 million records now after reading your article i want o use partition.I have two date columnn in this table one is Recd_Date and another is Entered_Date.Now i want to create partition based on entered date and that should be monthly.I am afraid to do that becauase this table is currently using for production.

And i do not have any file group in the database so needyour suggestion on this.

Sanjay

If you dont know exactly how many records will be inserting into partioned table then you cant say first 10 records,then i would like to menion % instead of fix numbers.
Please can you reply with answere.

Imran

if i create new coulums and new sql query (select, insert, update or delete) after i have do partition, Should i do partition again. Or my partition will see my new coulums and my new sql query. Also i can create new table too.

Thanks

pleases replay on mail id.

Thanks
Shital

then it is just inserting one row (if tbl_Prev_Education has one row)
when tbl_Prev_Education has more then one row then it is not inserting in to tbl_Education table

anyone plz help me

I have a requirement where a table on its own is 600GB and the data is kept for 30 days. Every weekend there is a job which triggers 40 others bulk insert jobs which pumps data into the database from different locations.
Since there is no common factor to create the partition I have decided to create it based on day
example
Every months has got 1,2,3,4……….30 or 31
so through of creating 31 file groups and attached 31 file ndf files.

I am not sure if this can be done as I tried couple of times but the partition cannot pick up values from the getdate() field…

Any help would be really appropriated!

Thanks
Brijesh

I have a very big database, so it takes me about 1 hours to run one report. Can I use partion to divide my database to many smaller partition by every month? Can we store the data of one month in one partition?

Example:

partition_2009_01: will store all data which insert in January.
partition_2009_02: will store all data which insert in February.
……

Thanks and Regards.

Will partitioning improve performance drastically and if yes how many partitions would be most suitable to have on these many number of records.

To Start, I would say…

Create new datafiles in same or different hard disk to improve IO Performance ( if it is raid 5 or 10 then for DR ).

Again each datafile can have multiple file groups, you can have similar objects created under one filegroup, like all tables in one filegroup and all Indexes in other file group ( this gives you oppurtunity to take backup of individual filegroup, also improves performance ..

Thanks

My Database is a Decision Support System.

I have table that contains 100 Million records, and each day we bulk copy and insert 300,000 more records. The main problem we are facing is as per our logic, the data for the past 25 days is only required when performing other business logic. We do not want to have archiving and moving old data to history tables, so any given day all the data for the last 2 years exist. Now my problem is when creating indexes I dont need the index to be on the previous months rather than I need the index if it can cover the last 1 month of Date range. If I do table partitioning how will be my performance increase or is there any other way round for this problem.

you have partitioned for new table but how do we know this table gets filled in future? I am not clear this point (New table partitioning) can u clear me out this?

Thanks in advance

i have one doubt in the post. Here ‘TestDB_Part2.ndf’ is ‘.mdf’ or ‘.ndf’?

Did u succeed in your project , Can you please share your knowledge ?

This article is excellant. Thanks for this.
I have a very huge table and i wish to partition this table based on a date field with out affecting other tables. How can i partition this table which has millions of records? Any impacts on partitioning a table which has data?

Thanks In Advance,
Anoop

We are facing some problems in our company.

we have a very huge dataset ~100M and we have used partitioning for the same.
All the data was supposed to be residing in the partitions created according to the statecodes but the size of primary mdf file has increased to a bulk.
Can you suggest a solution or reason for.

Regards
Ajay

CREATE PARTITION FUNCTION ParitioningTest_PartitionRange(nchar(4))
AS
RANGE LEFT FOR VALUES (‘CARD’, ‘CBO’, ‘CHEQ’, ‘CUST’, ‘TELL’)

CREATE PARTITION SCHEME ParitioningTest_PartitionScheme
AS PARTITION ParitioningTest_PartitionRange
TO ([FG_CARD], [FG_CBO],
[FG_CHEQ],[FG_CUST],
[FG_TELL], [FG]);

Now I dont know how to move records from ‘Transaction_Summary’ table to this partitions.

I should mention that all this partitioning is for this “Transaction_Summary” which has more than 100 million records and there in a CHAR(4) field in this table called “Channel” which doesnt have an index(there is an id column which is primary key” and I want to do the partitioning based on this filed, I know if that filed was a primary ke I could say

Drop index Transaction_Summary_IXC on Transaction_Summary with (Move To [ParitioningTest_PartitionScheme] (Channel) )

But I dont know what to do in this case, could you please help me?

Regards,
Hooman