Comments on: Index Efficiency http://jonathanlewis.wordpress.com/2008/05/14/index-efficiency/ Just another Oracle weblog Sat, 18 May 2013 11:04:10 +0000 hourly 1 http://wordpress.com/ By: Index Efficiency 2 « Oracle Scratchpad http://jonathanlewis.wordpress.com/2008/05/14/index-efficiency/#comment-35643 Sun, 28 Feb 2010 15:40:02 +0000 http://jonathanlewis.wordpress.com/?p=367#comment-35643 [...] Troubleshooting — Jonathan Lewis @ 3:39 pm UTC Feb 28,2010 Some time ago I wrote a note on index efficiency that referenced, in one of the comments, an article on my old website about how you could use [...]

]]> By: Jonathan Lewis http://jonathanlewis.wordpress.com/2008/05/14/index-efficiency/#comment-32842 Sat, 25 Apr 2009 11:21:32 +0000 http://jonathanlewis.wordpress.com/?p=367#comment-32842 Hans-Peter,
I’ve recently made use of the script that I’ve referenced above – and realised that it has two significant errors. I’ve pointed them out on the original posting, and plan to write up a more useful example on my blog some time in the next few weeks.

]]> By: Index analysis « Oracle Scratchpad http://jonathanlewis.wordpress.com/2008/05/14/index-efficiency/#comment-31910 Mon, 29 Sep 2008 11:58:21 +0000 http://jonathanlewis.wordpress.com/?p=367#comment-31910 [...] Index Rebuild 10g [...]

]]> By: Jonathan Lewis http://jonathanlewis.wordpress.com/2008/05/14/index-efficiency/#comment-31081 Thu, 22 May 2008 10:08:32 +0000 http://jonathanlewis.wordpress.com/?p=367#comment-31081 Hans-Peter,

It looks like we’ve had similar thoughts in the past. There is a second post on my website that describes that strategy and supplies some SQL for simple B-tree indexes.

]]> By: Hans-Peter Sloot http://jonathanlewis.wordpress.com/2008/05/14/index-efficiency/#comment-31080 Thu, 22 May 2008 08:31:03 +0000 http://jonathanlewis.wordpress.com/?p=367#comment-31080 Hello Jonathan,

What about the following calculation.
Sum the average column length of all columns that make up the index.
Add 10 for the rowid to this summed column length.
Multiply this with a factor to account for overhead (I have taken 1.3).
Multiply this by the number of rows from dba_indexes this would give us the number of bytes needed for the leaf blocks.
Compare this to the real number of leaf blocks.

This does not account for the fact that columns can have null values in concatenated indexes.

Regards Hans-Peter

]]> By: joel garry http://jonathanlewis.wordpress.com/2008/05/14/index-efficiency/#comment-31057 Thu, 15 May 2008 21:09:15 +0000 http://jonathanlewis.wordpress.com/?p=367#comment-31057 Still much appreciated. I’ll be importing this data into new hardware/app version/O10g in a month and a half, but I don’t expect the app will do anything different there regarding these indices. In case anyone wondered, I haven’t rebuilt them, they’ll be imp’d soon enough, and I’ll have a lot more space anyways (uh oh, sounds like famous last words).

]]> By: lascoltodelvenerdi http://jonathanlewis.wordpress.com/2008/05/14/index-efficiency/#comment-31052 Thu, 15 May 2008 12:32:13 +0000 http://jonathanlewis.wordpress.com/?p=367#comment-31052 I was looking for this!

Thant you very much.

]]>