I wouldn't start splitting out your functionality into separate sprocs in the hope it will improve performance. Spend the time to understand why the slow ones are slow,...

Sorry, I misunderstood your question, and ignored the issue of sequence.

I leave this to the actual experts.

Provided the "missing" values are represented by NULLs in your db, something like this will identify those rows:

select company_id, Unit_ID, Device_id, TX_DT_TM from theTable where TX_NUM is NULL

A quick'n'dirty 2 stage solution:

(1)

This gives you the table names that contain the column

select distinct sys.tables.name from sys.tables, sys.columns

where sys.tables.type = 'u'

and sys.tables.name not like ...

2³² = 4GB

A processor can only use memory of which...

2³² = 4,294,967,296 = 4GB

try enclosing the dates in quotes:

WHERE RegistrationDate between '2010-01-01' and '2010-03-31'

This looks problematic:

DECLARE @ResumePrefix VARCHAR(50)

DECLARE @ResumeSuffix VARCHAR(50)

DECLARE @ResumeLastNo BIGINT

DECLARE @uniqueno varchar(50)

SELECT @uniqueno = ResumePrefix+ResumeSuffix+ResumeLastNo

Oracle's recent marketing slogan certainly beats anything Microsoft every cooked up:

"ORACLE...Software...Hardware...Computers."

Can't wait till CocaCola gets on that bandwagon

"COKE ...Pop...Can...Beverages."

Run this:

Select 1/Nullif(0,0)

Then run this:

Select 1/count(Nullif(0,0))

This illustrates the situ.

select count(Nullif(0,0))

I believe your use of NULLIF(...) is valid here.

It's the call to COUNT(...) that can return a 0.

Why have that call to COUNT() there?

If F.IsSubjectToApproval = 0, then this call:

...COUNT(Nullif(F.IsSubjectToApproval,0)...

Returns 0 ...KABLOOIE!

No apologies neccessary Andy. This is the 1st time all week I did something right.:blink:

Looks good to me,...

Why?