Home » SQL & PL/SQL » SQL & PL/SQL » Clarification on 'N' th High/Low salary
Clarification on 'N' th High/Low salary [message #232770] Mon, 23 April 2007 04:16 Go to next message
Messages: 23
Registered: April 2007
Junior Member

Apologies at the outset for trying the thread for the umpttenth time. I just need a clarification regarding usage criteria here.
I can use the RANK function here and there are a zillion other ways to tackle this.
And I can use a plain old fashioned query like

select o.empno,o.ename,o.sal 
from emp o
where 3 > (select count(distinct(sal)) from emp where o.sal <sal)
order by sal desc

I assume I don't need to give the EMP table details
I want to know since I don't get any significant cost/time benefit why should I use the RANK function when I can achieve the same? Or is the standard(read elegant) way to do things

Apologies again for the trouble.

Re: Clarification on 'N' th High/Low salary [message #232787 is a reply to message #232770] Mon, 23 April 2007 05:00 Go to previous message
Michel Cadot
Messages: 63923
Registered: March 2007
Location: Nanterre, France, http://...
Senior Member
Account Moderator
Read your query.
You scan emp table twice.
Using analytic functions, you scan it only once.

This is the reason to use rank.

Previous Topic: like function
Next Topic: Indexes - Disdavantages
Goto Forum:

Current Time: Tue Oct 25 12:49:07 CDT 2016

Total time taken to generate the page: 0.11707 seconds