select convert(bigint,HashBytes('SHA1', convert(varchar, getdate(),121)+convert(varchar,rand())))

this should be a bit unique:cool:

He said that all columns are varchar, so "coloumn * 1" would just convert the varchar to an int/float then multiply by 1, giving the int value. Better off just...

SSIS can preview the file just fine and I opened it up in notepad++ and didn't see any issues with visible and special chars

you want 8.00 instead of 8.009. So you go from 3 spaces after to 2 spaces and...

if that did it, then hopeflly those Seq's keep incrementing. Also, those are 3 distinct selection statements going on, not the best.

I may have misunderstood something, but what about this?

create table #T (

Seq int primary key,

Account int,

Disp char)

--drop table #t

insert into #T (Seq, Account, Disp)

select 10,12345,'+' union all

select 20,23451,'' union all

select...

I haven't had to work with dynamic SQL yet, but I've heard some people mention that the execution plan can get messed up some times. You may want to see...

I don't work with multiple servers, but I heard of having different DB servers incrementing on different values.

eg..

db1 goes: 1,4,7.....

db2 goes: 2,5,8.....

db3 goes 3,6,9.....

Then you never have duplicate keys. But...

Sounds like MOST rows are being returned. An Index usually only helps if rows are ~95+% unique.

Anytime you're returning a ton of rows...

It's nice to know...

Here's a question. Would overall performance be better when joining on a dept id vs name because of the comparing of int vs string? I would assuming other tables store...

If Google can't find it, it isn't real... 😛

on a side note, if everything you're comparing/returning is 'covered' by the...

I've seen many times where MS said they plan on post Windows Server 2008 and SQL 2008 to be 64bit only. If you don't have a 64bit server by 2011,...

I found "UPDATE STATISTICS ". Is there any way to script the command run on my raw tables parallel to the live table import or would I have to...

You can pretty much guaranteed your statistics are garbage when truncating and reloading.

Is your source data sorted according to the clustered index? If not you may be better off...