Comments on: The power of T-SQL’s APPLY operator

By: Rob Farley

Rob Farley — Thu, 01 Mar 2018 14:12:50 +0000

Chris – try grabbing it from https://1drv.ms/v/s!AoK2mL8SvaMytzvfE5N1LOrdIJgj

By: Rob Farley

Rob Farley — Thu, 01 Mar 2018 14:08:36 +0000

Hi Chris,

I don’t have much control over the SQLBits site. I can try to ask them maybe…

But I’m not sure I do cover the idea of using a view instead of APPLY. Both are useful in their way, but APPLY is certainly very useful for the fact that it provides correlation in a much easier way.

By: Chris

Chris — Mon, 05 Feb 2018 15:26:10 +0000

I’m intrigued by the aside about using begin/end rendering the sql “non-simplifiable.” Unfortunately, while the shortcut URL http://bit.ly/SimpleRob correctly links to http://sqlbits.com/Sessions/Event5/Designing_for_simplification , the presentation on the destination page is not functioning, nor is the link to the presentation wmv file (the links to ppt and demo sql work fine, though).

Specifically, I’m not seeing why APPLY with a table function is simpler than – or preferable to – an inline view. Most correlated subqueries I’ve come across can be converted to a join to an inline view in the main body of the procedure without using a correlation in the where clause. This is likely covered in the broken presentation link.

By: Jodie

Jodie — Tue, 28 Jun 2016 18:16:40 +0000

I know I’m a few years late, but just wanted to say thanks for such a well-explained article and the highly useful examples. I will definitely keep Apply in mind for future SQL conundrums! Thank you 🙂

By: Johnny

Johnny — Sat, 03 Oct 2015 01:25:49 +0000

Great article.

By: Jean Joseph

Jean Joseph — Thu, 16 Oct 2014 03:04:24 +0000

Cool article

By: Valeriy Khvalabov

Valeriy Khvalabov — Fri, 13 Jun 2014 17:46:31 +0000

Thank you Rob,
A great article!
You made the explanation simple and straight-forward.

By: Rob Farley

Rob Farley — Fri, 06 Jun 2014 05:13:09 +0000

You can always use TOP (3) WITH TIES…

By: Andrew Watson

Andrew Watson — Fri, 06 Jun 2014 04:54:48 +0000

One minor comment.. While top(3) works… I would humbly suggest that it won’t cope very well if the Business requirement, is "return the top 3 performers, by sales amount"… if the amounts are ever tied….
eg:if the product sums are:
One:1000
Two: 800
Three: 500
Four: 500
Would the business want to return the top 4 in this case (as there is a tie for third place), or have a random selection of Three/Four?.
If they would want to see both, then the With query with a RANK (rather than Row_number) function is the more appropriate choice…. definitely more adjustable to business rules (rank = top performers by amount, dense_rank = top amounts and products with those amounts, row_number = top three by amount [and some other ordering criteria])… don’t fight it. 🙂

By: Dave

Dave — Mon, 10 Mar 2014 08:51:01 +0000

…and it works fine doh. Poor internet today