gvoshol 73146 (9/6/2016)

---

If you use an identity, there is nothing inherently prohibiting you from entering two duplicate records except for the key. In the example of the state table,...

I'd say the level of fragmentation indicates that there is a lot of split pages occurring on inserting/updating data.

Those splits...

Rethink your approach.

It's better to have a separate "ChangeAudit" table where you record all the changes done to the tables.

Normally 1 table to store changes to all audited tables is...

Loop back is not so much of a problem as distributed transactions which hold exclusive lock on remote servers.

Try to avoid pushing data to a remote server, it always better...

There is a much simpler way to find time difference between 2 datetime values:

TimeDifference = TimeTo - TimeFrom

THis will allow to count milliseconds too:

;WITH DEMO(ID,FromTime,ToTime) AS

(

SELECT...

Those are completely non-related things.

Clustered index does not have...

Here are some reasons to consider.

A PK on a "popular" look-up table has to be mentioned in every other table referencing it.

If that PK has 3 or more columns -...

Procedure B could simply read records from Table A

Add time to the statistics you capture:

Set Statistics IO,TIME On

It will tell you why.

Columns are checked. If a columns does not exist sp_get_query_template throws an error.

Tables...

The fault in the query must be too precious for your company.

Don't you dare to fix it!

:hehe:

You may replace actual column names with Col1, Col2, etc.

And post not the whole query, but only that part where the tables involved in the join are mentioned.

Again, using made-up...

Sergiy (8/31/2016)

---

I'll have to disagree with you there, ol' friend. Automatic truncation to a scale of 6 isn't correct. At least not in my book. If a dollar store calculator...

I can think of one - fighting parameter sniffing...

Please post the join and DDL for both tables involved.