June 6, 2013 at 5:11 pm
Hi team,
I am looking for Partitioning / Slip the Data in a table to Multiple Small Files.
Logical Partition can be done based on a Varchar Column WHich store currency code
EX:
SELECT CUR_CODE, COUNT(*) FROM TEST GROUP BY CUR_CODE
O/P:
GBP - 10000
USD - 20000
SGD - 30000
But when I try to create partition, on providing the Split Criteria and details and hit Estimate storage
It show the details in this way
GBP 1
USD 50000
SGD 1
Please need input to split the data based on the Value in the COlumn, Expecting to Send the Data to Each Currency File
GBP.ndf - 10000
USD.ndf - 20000
SGD.ndf - 30000
Thanks in advance
June 6, 2013 at 5:42 pm
I just want to clarify, you're looking to setup table partitioning within the schema of the database to multiple filegroups, correct? The reason I'm trying to clarify is because I've never seen an estimate for storage before, are you using a wizard?
Additionally, is there an index on the field you're looking to partition across? I'm wondering if the statistics are just wildly off. If there's an index on that field, UPDATE STATISTICS WITH FULLSCAN the table and see what kind of estimates you get.
Finally, and I hope you're on a dev system, just chuck the estimates out and do your partition. Then go to the dm_ views and check your partition sizes with actuals. Estimates are notorious for being wrong.
Never stop learning, even if it hurts. Ego bruises are practically mandatory as you learn unless you've never risked enough to make a mistake.
For better assistance in answering your questions[/url] | Forum Netiquette
For index/tuning help, follow these directions.[/url] |Tally Tables[/url]
Twitter: @AnyWayDBA
June 8, 2013 at 3:11 pm
I just want to clarify, you're looking to setup table partitioning within the schema of the database to multiple filegroups, correct?
> Yes I am trying to Partition my table, here. with multiple file group and multiple files.
Ex: USD will have USD_FG and USD_FILE.ndf file
The reason I'm trying to clarify is because I've never seen an estimate for storage before, are you using a wizard?
> Yes I was using Partition based on Wizard
Additionally, is there an index on the field you're looking to partition across? I'm wondering if the statistics are just wildly off. If there's an index on that field, UPDATE STATISTICS WITH FULLSCAN the table and see what kind of estimates you get.
> No Indexed are on the table
Finally, and I hope you're on a dev system, just chuck the estimates out and do your partition. Then go to the dm_ views and check your partition sizes with actuals. Estimates are notorious for being wrong.
> Yes Im on a test DB, I will Test it.
June 10, 2013 at 1:11 pm
sudhirnune (6/8/2013)
Additionally, is there an index on the field you're looking to partition across? I'm wondering if the statistics are just wildly off. If there's an index on that field, UPDATE STATISTICS WITH FULLSCAN the table and see what kind of estimates you get.> No Indexed are on the table
Sorry, fell off the earth a bit on this. You have no indexes? No wonder you can't get a good estimate. If this table is a heap you're not going to get any performance increases from partitioning, you're aware of that, right?
Partitioning is more an administrative tool to ease data manipulation, not an optimization tool.
Never stop learning, even if it hurts. Ego bruises are practically mandatory as you learn unless you've never risked enough to make a mistake.
For better assistance in answering your questions[/url] | Forum Netiquette
For index/tuning help, follow these directions.[/url] |Tally Tables[/url]
Twitter: @AnyWayDBA
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply