Great... More missing posts... :crazy:

If anyone w/ admin privileges can find my last post on this thread (index test w/ results), please de-spam it... Because, I'm not rewriting it...

Testing the assumption that a covering index would negate the impact of the sort operator cause by ordering on an expression (rowrank).

The test table... Just under 1/2 M rows similar...

j-1064772 (5/29/2015)

---

Eirikur Eiriksson (5/25/2015)

---

[font="Comic Sans MS"]Sinking feeling of reverting back to...

Alan.B (5/29/2015)

---

Yeah, unfortunately make_parallel can't help you with that:-P

I have a couple functions for digging out data sets, parameter info and stuff like that. I'll post them here when I...

Well now... Turns out that the SQL Server instance that's running Reporting Services only has a single CPU core available to it...

I suppose that's why a query that has an...

Alan.B (5/29/2015)

---

I actually did a presentation that covered this very topic at a recent BI user group but don't have the code handy at the moment.

My experience has been...

Jeff Moden (5/28/2015)

---

Jason,

It looks like the auto SPAM post hider got ya. I can see all of your posts but I can't make them visible for you. Not...

The only part you're missing is a means to maintain the hierarchy/order once the data is inserted into the table.

You'll noticed that I added an IDENTITY column to the...

Jeff Moden (5/28/2015)

---

Jason A. Long (5/28/2015)

---

I would suspect that the answer to which one is faster "quirky method" or "windowed function", is going to be heavily dependent of the available...

Jeff Moden (5/28/2015)

---

Jason A. Long (5/28/2015)

---

I would suspect that the answer to which one is faster "quirky method" or "windowed function", is going to be heavily dependent of the available...

Jeff Moden (5/28/2015)

---

Jason A. Long (5/28/2015)

---

I would suspect that the answer to which one is faster "quirky method" or "windowed function", is going to be heavily dependent of the available...

So... Just knocked out the 1st test...

I wanted to start off with a simple "Apples to Apples" race on a simple 3 column table with a single key clustered...

I would suspect that the answer to which one is faster "quirky method" or "windowed function", is going to be heavily dependent of the available indexes.

The "quirky method" relies exclusively...

The following should give you what you're looking for.

IF OBJECT_ID('tempdb..#temp') IS NOT NULL

DROP TABLE #temp;

CREATE TABLE #temp (

TN CHAR(4),

CallDate DATE

)

INSERT #temp (TN, CallDate) VALUES

('1111','2014-01-01'),

('1111','2014-02-01'),

('1111','2014-03-01'),

('1111','2014-06-01'),

('1111','2014-07-01'),

('1111','2014-10-01'),

('1111','2014-11-01'),

('1111','2014-12-01'),

('1111','2015-01-01'),

('1111','2015-02-01'),

('1111','2015-03-01'),

('1111','2015-04-01'),

('1111','2015-08-01'),

('1111','2015-09-01'),

('1111','2015-10-01'),

('1111','2015-11-01'),

('1111','2015-12-01');

SELECT

t1.TN,

t1.CallDate,

cc.CallCount

FROM

#temp t1

CROSS APPLY (

SELECT COUNT(*)...

Use the file I supplied as a guide and try to recreate it on your own....