I did. Mine was faster 😉

Jeff Moden (3/16/2009)

---

Do you have the code for the new function... I...

I can't seem to get better performance than my original function.

I'm pasting it here for future reference.

[font="Courier New"]SET ANSI_NULLS ON

SET QUOTED_IDENTIFIER ON

GO

ALTER FUNCTION [dbo].[GetWordCount]

(@HTMLText VARCHAR(MAX))

RETURNS INT

AS

BEGIN

DECLARE @Start  INT

DECLARE @End    INT

DECLARE @Length...

Jeff Moden (3/16/2009)

---

One of your requirements was to ....

4. Replace all entity codes (“&xxx;”) with X

If you "smear" that requirement across all the...

For script & style tags I need to remove everything they contain.

For other HTML tags I can just remove the tags...

I have to remove script tags...

Jeff, that looks awesome.

I'm going to test it out at work tomorrow.

One part I don't quite understand is this:

--===== Replace all EntityCodes with the actual ascii character (Just to...

RBarryYoung (3/15/2009)

---

Goldie: in Rule 2 (Step 1) what character are you replacing with a space? All of my attempts to extract/examine it from your post show it as a...

Download the file and save as text.

If you try to view it in your browser it will translate the HTML tags.

Jeff Moden (3/14/2009)

Hi Goldie...

Not sure I can help but it would make it easier to try if you attached a file that has an example HTML that you're trying...

Bob Hovious (3/13/2009)

---

Thoughts, since you have some logic laid out already.

First, I see you know how to use tally tables, so use them ... over and over again... get rid...

Jack Corbett (3/13/2009)

---

I have to say that this sounds like the ideal place to create a CLR function so you can use Regular Expressions to clean up the data. ...

If you post some sample DDL & DATA we would be happy to help you solve your problem.

Please read this article for more information on how to get the best...

Jeff Moden (1/31/2009)

---

Nah... I take it back... not clever... freakin' brilliant. The key to this whole thing is the grouping. You've gotta be able to group each "cluster"...

You could also try a dynamic cross-tab, as mentioned in the following article:

http://www.sqlservercentral.com/articles/cross+tab/65048/