Those aren't datasets, they are CTE's, which are somewhat like views in that they are more or less...

ashishkumarrai (9/6/2016)

---

Hi

Thanks for the post, really useful. I am working on similar stuff. My problem is the value the above query is using is static. What will be the...

Y.B. (9/6/2016)

---

ChrisM@Work (9/6/2016)

---

Sean Lange (9/6/2016)

---

Luis Cazares (9/6/2016)

---

Sean Lange (9/6/2016)

---

I tried my hand at curing my own bacon for the first time. It came out of the cure Saturday morning and...

Sean Lange (9/6/2016)

---

Luis Cazares (9/6/2016)

---

Sean Lange (9/6/2016)

---

I tried my hand at curing my own bacon for the first time. It came out of the cure Saturday morning and hit the...

The Dixie Flatline (9/5/2016)

---

We do something similar, matching on license numbers...

That's a fair description of what the code does -...

Try this fuzzy-matching function:

/****** Object: UserDefinedFunction [dbo].[IF_Levenshtein02] Script Date: 27/01/2014 20:12:53 ******/

SET ANSI_NULLS ON

GO

SET QUOTED_IDENTIFIER ON

GO

-- this will score around 10,000 word pairs per second on...

The Dixie Flatline (8/12/2016)

---

An alternative is to just back up the starting date to...

The Dixie Flatline (9/2/2016)

---

What should the result be if a row matches two or more established groups? For example, there is a 'Smith' group and a 'Jones'...

ScottPletcher (9/2/2016)

---

If the overall number of pages to be read is small either way, it's just not worth the optimizer's time to keep going.

SQL naturally favors the clustered index. ...

GilaMonster (9/2/2016)

---

Abort reason is "Good enough plan found" (properties of the SELECT). Optimisation has multiple phases, "Good enough plan found" means that at the end of one phase the cheapest...

raymond.wee.823 (9/2/2016)

---

Hi, it looked promising. However, I tried it with 10000 records and it took 10 minutes just to go through the first 1000 records so it is a performance...

raymond.wee.823 (9/2/2016)

---

Hi, it looked promising. However, I tried it with 10000 records and it took 10 minutes just to go through the first 1000 records so it is a performance...

Here's a slightly different version, replacing table reads with table spools:

WITH

Firstpass AS (SELECT ID, [Name], Phone, Email, [Group],

grp1 = MIN(ID) OVER(PARTITION BY [Name]) FROM #Temp),

Secondpass AS (SELECT ID,...

-- Set up some sample data to test against

-- This is not part of the solution

-- Note: the ID column is required, you could generate it on-the-fly using ROW_NUMBER()

SELECT ID,...