I think those are also uniqueidentifier columns, so fragmentation is to be expected. Unless you retain lots of backup info in msdb, and I would not recommend that, then...

I think those are also uniqueidentifier columns, so fragmentation is to be expected. Unless you retain lots of backup info in msdb -- and I would not recommend that...

You need to review the choice of clustering key; the clustering key is critical to performance, so never be lazy enough to allow the clustering key to default to identity....

Yes, the index row length will be 20 bytes. That is normal. Since bigint is a fixed-size data type, even if the value is NULL, it will still...

dwain.c (4/8/2015)

---

ScottPletcher (4/8/2015)

---

I'm not seeing why you can't just use the code below. I'm sure someone can quickly explain the part of the problem that I'm missing :-).

SELECT cust_id,...

I'm not seeing why you can't just use the code below. I'm sure someone can quickly explain the part of the problem that I'm missing :-).

SELECT cust_id,

...

I'd expect value -127 to take 2 bytes, since smallint is the smallest data type that can hold that value. Thus, it would still save 2 bytes vs. an...

You need to tell SQL how to match the multiple values. A couple of quick examples:

Set @KeywordSearch = 'Platform AND Customer AND Support'

Set @KeywordSearch = 'Platform NEAR Customer NEAR...

I would hope combining them into a single transaction with a held lock would prevent that problem:

begin transaction

select @result = Result

from Cache with (rowlock,updlock)

where [Hash] = @hash

-- A

if (@result...

If RCSI is on, you wouldn't have any need to use ALLOW_SNAPSHOT_ISOLATION.

This would be my only suggestion for an alternative. Don't think it's any better at all, other than that it can handle an input value this is only 4...

Just to be sure to get the best lookup, insert the individual values in the IN list into a table uniquely clustered on the value. Then, again just to...

Sorry, I'm a DBA, so I'm highly technical on those terms ;-).

real.forte 18467 (4/3/2015)

---

Thank you very much for the information. Here is what I ended up doing:

1. I fetched the SID of the dbo on the original server

2. Provided I...

CREATE PROCEDURE DvzkuruCek

@T1 DATETIME = '20150202' --'YYYYMMDD' is always interpreted correctly

AS

SET NOCOUNT ON;

DECLARE

@KUR INT;

SELECT TOP (1) @KUR = RATES1

FROM dbo.L_DAILYEXCHANGES

WHERE EDATE= @T1 AND CRTYPE...