Comments on: How simple answer are you looking for ? http://www.mysqlperformanceblog.com/2007/04/11/how-simple-answer-are-you-looking-for/ Everything about MySQL Performance Tue, 02 Dec 2008 13:39:19 +0000 http://wordpress.org/?v=2.5.1 By: peter http://www.mysqlperformanceblog.com/2007/04/11/how-simple-answer-are-you-looking-for/#comment-115363 peter Sat, 28 Apr 2007 17:45:16 +0000 http://www.mysqlperformanceblog.com/2007/04/11/how-simple-answer-are-you-looking-for/#comment-115363 Bob, Please ask questions which are not related to the topic in our forums instead. The answer to your question would be - creating view is not going to help as view does not carry any of its own indexes or other special structures. Bob,

Please ask questions which are not related to the topic in our forums instead.

The answer to your question would be - creating view is not going to help as view does not carry any of its own indexes or other special structures.

]]> By: Bob Bankay http://www.mysqlperformanceblog.com/2007/04/11/how-simple-answer-are-you-looking-for/#comment-114543 Bob Bankay Thu, 26 Apr 2007 21:30:26 +0000 http://www.mysqlperformanceblog.com/2007/04/11/how-simple-answer-are-you-looking-for/#comment-114543 We have a large table (10M row) with about 12 indices. One of then indicates that state of the row in the table and there are about 30 states. The states change from 0 to 30 in about 14 days. Would it be appropriate to create views for state with the highest population. Would it be faster than a lookup for the specific state wiht LIMIT n? We have a large table (10M row) with about 12 indices. One of then indicates that state of the row in the table and there are about 30 states. The states change from 0 to 30 in about 14 days.

Would it be appropriate to create views for state with the highest population. Would it be faster than a lookup for the specific state wiht LIMIT n?

]]> By: promotingblogs.com http://www.mysqlperformanceblog.com/2007/04/11/how-simple-answer-are-you-looking-for/#comment-107739 promotingblogs.com Fri, 13 Apr 2007 14:49:22 +0000 http://www.mysqlperformanceblog.com/2007/04/11/how-simple-answer-are-you-looking-for/#comment-107739 <strong>MySQL Performance: interesting tip about indexes...</strong> I’ve got an interesting comment the other day saying “I’ve heard Jay Pipes saying indexes with cardinality lower than 30% are worthless, true?” That is interesting question and it has different answers depending on who is asking. A lot of peopl... MySQL Performance: interesting tip about indexes…

I’ve got an interesting comment the other day saying “I’ve heard Jay Pipes saying indexes with cardinality lower than 30% are worthless, true?” That is interesting question and it has different answers depending on who is asking. A lot of peopl…

]]> By: peter http://www.mysqlperformanceblog.com/2007/04/11/how-simple-answer-are-you-looking-for/#comment-106709 peter Wed, 11 Apr 2007 21:10:50 +0000 http://www.mysqlperformanceblog.com/2007/04/11/how-simple-answer-are-you-looking-for/#comment-106709 Jay, I'm not saying you're saying that. I'm simply quoting the guy posted comment on this blog :) I just thought that would be good way to start the post being nice teaser. The point is not what you're giving people simple answers to complex questions but what a lot of people are looking for simple answers because it makes them comfortable. Jay,

I’m not saying you’re saying that. I’m simply quoting the guy posted comment on this blog :)

I just thought that would be good way to start the post being nice teaser.

The point is not what you’re giving people simple answers to complex questions but what a lot of people are looking for simple answers because it makes them comfortable.

]]> By: Jay Pipes http://www.mysqlperformanceblog.com/2007/04/11/how-simple-answer-are-you-looking-for/#comment-106560 Jay Pipes Wed, 11 Apr 2007 15:10:41 +0000 http://www.mysqlperformanceblog.com/2007/04/11/how-simple-answer-are-you-looking-for/#comment-106560 Uh, I would never say "indexes with cardinality lower than 30% are worthless". ;) What I have said in the past is that if you have a situation where the value you are searching for is contained in 30% of an index's records, it is highly unlikely the index will be used, as the efficiency of a sequential scan will outweigh the numerous bookmark lookup operations needed with the index seek operations. Of course, having a covering index can influence the optimizer's decision of course, as will having multiple levels of cardinality (multiple columns in the index). Bottom line: I *never* say there is a simple answer or a single answer to anything. Just take a look at my latest set of slides... on slide 2, I say "The answer to every question is... it depends." Cheers, and see you soon at the conference, Peter! Uh, I would never say “indexes with cardinality lower than 30% are worthless”. ;) What I have said in the past is that if you have a situation where the value you are searching for is contained in 30% of an index’s records, it is highly unlikely the index will be used, as the efficiency of a sequential scan will outweigh the numerous bookmark lookup operations needed with the index seek operations. Of course, having a covering index can influence the optimizer’s decision of course, as will having multiple levels of cardinality (multiple columns in the index). Bottom line: I *never* say there is a simple answer or a single answer to anything. Just take a look at my latest set of slides… on slide 2, I say “The answer to every question is… it depends.”

Cheers, and see you soon at the conference, Peter!

]]>