select without order by and clustered index

답변

• 2008년 11월 18일 화요일 오후 11:42

  중재자

   

   
  

  0

  로그인하여 투표
  The Beatles versus the Stones

  http://sqlblog.com/blogs/hugo_kornelis/archive/2006/12/31/Beatles-vs-Stones.aspx

   

  Quaere Verum - Clustered Index Scans - Part I

  http://www.sqlmag.com/Article/ArticleID/92886/sql_server_92886.html

   

  Quaere Verum - Clustered Index Scans - Part II

  http://www.sqlmag.com/Article/ArticleID/92887/sql_server_92887.html

   

  Quaere Verum - Clustered Index Scans - Part III

  http://www.sqlmag.com/Article/ArticleID/92888/sql_server_92888.html

   

   

  AMB

  • 응답
  • 인용

   
• 2008년 11월 19일 수요일 오전 8:48

  답변자

   

   
  

  0

  로그인하여 투표

  Absolutely not.

  Here's an example I wrote a while back to prove this to a colleague

  Code Snippet

  CREATE TABLE dbo.order_by (
     id int identity(1,1) PRIMARY KEY CLUSTERED
   , some_field char(1)
  )

   

  INSERT INTO dbo.order_by (some_field) VALUES ('B')
  INSERT INTO dbo.order_by (some_field) VALUES ('A')

   

  SELECT *
  FROM   dbo.order_by

   

  CREATE NONCLUSTERED INDEX ni_order_by_some_field ON dbo.order_by (some_field ASC)

   

  SELECT *
  FROM   dbo.order_by

   

  GO
  DROP TABLE dbo.order_by

   

   

  Thanks for the other links hunchback, I will bookmark them as further ammunition

  • 응답
  • 인용

   
• 2008년 11월 19일 수요일 오후 2:37

  중재자

   

   
  

  0

  로그인하여 투표
  No.   If you do not use "Order by" the order of the result is "undefined".  Which means, it can be anything depending on how the server feels at the time.
  
  
  
  • 응답
  • 인용