Comparing the latest two execution plans, all the filtered index did is drop approximately 700,000 rows from the initial pull. May have reduced the reads, but not sure if...

As it looks like you are using SQL Server 2008 perhaps a filtered index. Not sure if it will help, but it is worth testing.

CREATE NONCLUSTERED INDEX [AI_OperatorCBGEstimate_14947]

...

You really haven't provided enough detail to really help you solve the problem you seem to be having. To help us help you we need the DDL (CREATE TABLE...

Looking at the latest execution plan, you are still reading over 14 million rows from the table that meet your search criteria. That is a lot of reads.

If it isn't a problem I wouldn't worry about it.

Just mho.

Doesn't answer the question. Is there a problem due to the gaps as a result of the deletes or do you want to do this for cosmetic reasons?

Just curious, but could you post the actual code you use to complete the counts?

My question, why do you need to eliminate gaps in the sequence numbering?

As Sean said, if you post the DDL (CREATE TABLE statement) for the table(s) involved, some sample data (a series of INSERT INTO statements (some people here still use SQL...

Try the same query using sys.objects.

If you are updating existing data you use the UPDATE statement. INSERT is used to insert new data.

Sean Lange (4/14/2014)

---

That does NOT help me understand what you want. I can't see your screen, I have no idea what datatypes are in your table, I have no idea...

I have to ask this question, what is the batch process doing?

This will work as well:

with SampleData as (

select

StrVal

from

(values ('abc ,'),('cde,xyz,aaa,'),('zzz, yyy ,ttt,'),('qwe'),('asd,fgh'))dt(StrVal)

)

select

StrVal,

left(StrVal,len(StrVal) - patindex(',%',reverse(StrVal)))

from

...

Two choices I can think of quickly, 1) figure out how to keep EF from converting the strings from varchar to nvarchar or 2) modify your table so that ObjId...