Comments on: Extending procedure_analyse http://fernandoipar.com/2009/04/21/extending-procedure_analyse/ I love mankind! Its people I can't stand! Wed, 27 Jan 2010 19:05:16 +0000 http://wordpress.org/?v=abc hourly 1 By: Fernando Ipar http://fernandoipar.com/2009/04/21/extending-procedure_analyse/comment-page-1/#comment-49 Fernando Ipar Wed, 22 Apr 2009 04:33:20 +0000 http://fernandoipar.com/?p=163#comment-49 Thanks for the additional suggestion, as I said, I expected that kind of problems. I'll look into that particular maatkit script for ideas, I don't intend to reinvent the wheel but rather to present DBAs with some useful screenshot of information, but it has to be useful and properly obtained :) Thanks for the additional suggestion, as I said, I expected that kind of problems. I’ll look into that particular maatkit script for ideas, I don’t intend to reinvent the wheel but rather to present DBAs with some useful screenshot of information, but it has to be useful and properly obtained :)

]]> By: Shlomi Noach http://fernandoipar.com/2009/04/21/extending-procedure_analyse/comment-page-1/#comment-48 Shlomi Noach Wed, 22 Apr 2009 03:58:54 +0000 http://fernandoipar.com/?p=163#comment-48 Hi, Still not complete. If the the index on (a) is UNIQUE, then an index (a,b) does make it redundant. Since there's already a utility which checks for duplicate indexes, and it's quite complicated, I suggest you take a look how it does it. It is part of maatkit, and is called mk-duplicate-key-checker. See <a href="http://code.google.com/p/maatkit/wiki/DeterminingDuplicateKeys" rel="nofollow">http://code.google.com/p/maatkit/wiki/DeterminingDuplicateKeys</a> Hi,
Still not complete. If the the index on (a) is UNIQUE, then an index (a,b) does make it redundant.
Since there’s already a utility which checks for duplicate indexes, and it’s quite complicated, I suggest you take a look how it does it. It is part of maatkit, and is called mk-duplicate-key-checker. See http://code.google.com/p/maatkit/wiki/DeterminingDuplicateKeys

]]> By: Fernando Ipar http://fernandoipar.com/2009/04/21/extending-procedure_analyse/comment-page-1/#comment-47 Fernando Ipar Wed, 22 Apr 2009 03:49:13 +0000 http://fernandoipar.com/?p=163#comment-47 Thanks a lot for the observation, that didn't occur to me :) I'll have to look just for columns that have count(*) > 1 and that are also indexed alone, this would be the useless index (of course, it might be that one index takes duplicates, the other not, I don't know if that still makes all cases useless, you've really opened a can of worms for me :) Fortunately we've got the thing at bitbucket now so I'll report this as a bug and begin working on it ASAP! Thanks a lot for the observation, that didn’t occur to me :)

I’ll have to look just for columns that have count(*) > 1 and that are also indexed alone, this would be the useless index (of course, it might be that one index takes duplicates, the other not, I don’t know if that still makes all cases useless, you’ve really opened a can of worms for me :)

Fortunately we’ve got the thing at bitbucket now so I’ll report this as a bug and begin working on it ASAP!

]]> By: Shlomi Noach http://fernandoipar.com/2009/04/21/extending-procedure_analyse/comment-page-1/#comment-46 Shlomi Noach Wed, 22 Apr 2009 03:22:51 +0000 http://fernandoipar.com/?p=163#comment-46 Good work. With regard to overused indexes, you should note that an index on (a,b) and on (a,c) have nothing useless between them. But your code will report "a" as being over indexed. Good work.

With regard to overused indexes, you should note that an index on (a,b) and on (a,c) have nothing useless between them. But your code will report “a” as being over indexed.

]]>