Viewing 15 posts - 271 through 285 (of 309 total)
Please note that it is not enough just to run the posted code.
You must also write your own code to split up the ipCountryLookup in 64K-sized blocks.
I just wanted to...
April 12, 2010 at 10:18 am
Jeff Moden (4/9/2010)
There's also a very high speed method that uses no character based conversions but I can't put my hands on it just now (still looking for it).
Please post...
April 12, 2010 at 8:54 am
If the performance is still not good enough you could preprocess the IpCountryLookup to split all blocks that are larger than 64K into several smaller blocks.
You could then rewrite the...
April 12, 2010 at 8:18 am
Some random tips:
1) Add a nonclustered index on CountryCode, IpNumber on [my].[ipAddresses]
I assume that when you run this update most of the rows in the table already have a CountryCode....
April 12, 2010 at 7:04 am
r5d4 (4/12/2010)
Anyone got any ideas to improve on this?
In what way do you want to improve it ?
Does it not work correctly ?
Performance ?
Disk space ?
Just asking us to...
April 12, 2010 at 6:05 am
I agree completely with Wayne.
/SG
April 9, 2010 at 5:18 am
Sure, 8 million rows should be no problem. Queries against such a table with an appropriate index will still be blindingly fast.
/SG
April 9, 2010 at 12:58 am
You should define a clustered index on GeoID, ProductID, AsOf just as you suggested yourself.
You should also change the query so the SQL server understands that you are really asking...
April 8, 2010 at 11:07 am
The only way to know how much compression you would actually get is to actually try it on your data.
Install SQL 2008 developer somewhere
Import a sample of your largest tables...
April 8, 2010 at 10:41 am
I would use a precomputed table of distances between all zipcodes and all stores.
Since you only have a few hundred stores this would give a much smaller table than the...
April 8, 2010 at 10:18 am
Judy-363345 (3/31/2010)
Table structure
CREATE TABLE [dbo].[mytable](
[ID] [bigint] IDENTITY(1,1) NOT NULL,
[UID] [uniqueidentifier] NOT NULL,
[ReferenceNumber] [varchar](20) NULL,
I am pretty sure that your problem is caused by allowing NULL in ReferenceNumber.
Try changing the table...
April 6, 2010 at 7:55 am
You probably have lots of overlapping queries. If your application generates 20 simultaneous queries that all start at 00:00:00 and then all finish at 00:00:02 the elapsed time would be...
March 30, 2010 at 3:27 am
Hmm. I see your point.
The problem is that if I do this with CREATE INDEX I will probably have to create new filegroups and new partitioning schemes using the new...
March 26, 2010 at 2:39 am
Paul, thank you for the idea of using EMPTYFILE to move all the data to a new file.
My current plan is to use backup/restore to move the data to...
March 25, 2010 at 5:01 pm
Kevin Bullen (3/25/2010)
March 25, 2010 at 10:02 am
Viewing 15 posts - 271 through 285 (of 309 total)