Comments on: What I hate about MySQL http://www.mad-computer-scientist.com/blog/2008/10/18/what-i-hate-about-mysql/ IT Professional by day, mad computer scientist by night Tue, 19 Jul 2011 17:26:05 +0000 hourly 1 http://wordpress.org/?v=3.2.1 By: mmcdermott http://www.mad-computer-scientist.com/blog/2008/10/18/what-i-hate-about-mysql/comment-page-1/#comment-921 mmcdermott Mon, 27 Oct 2008 16:23:17 +0000 http://mad-computer-scientist.com/blog/?p=80#comment-921 @pcdinh: I really can't say that I agree on constraints. Stored procedures I can see as a matter of taste, especially if one is using prepared statements, but not constraints. I have seen too much code where the constraints were not being enforced (either in MySQL or that they had never been created in the first place) and things broke. Rules that should have helped to ensure valid data were not being enforced and it caused a big mess. You may be able to somewhat get away with it when you're just building a website (after all, an orphaned article is an unfortunate bug, but not a catastrophe), but it is potentially huge when working with actual software. @pcdinh:

I really can’t say that I agree on constraints. Stored procedures I can see as a matter of taste, especially if one is using prepared statements, but not constraints. I have seen too much code where the constraints were not being enforced (either in MySQL or that they had never been created in the first place) and things broke. Rules that should have helped to ensure valid data were not being enforced and it caused a big mess. You may be able to somewhat get away with it when you’re just building a website (after all, an orphaned article is an unfortunate bug, but not a catastrophe), but it is potentially huge when working with actual software.

]]> By: mmcdermott http://www.mad-computer-scientist.com/blog/2008/10/18/what-i-hate-about-mysql/comment-page-1/#comment-920 mmcdermott Fri, 24 Oct 2008 15:23:03 +0000 http://mad-computer-scientist.com/blog/?p=80#comment-920 @Jonathan - right you are; I knew I overlooked a thing or two :) @Jonathan – right you are; I knew I overlooked a thing or two :)

]]> By: Matt http://www.mad-computer-scientist.com/blog/2008/10/18/what-i-hate-about-mysql/comment-page-1/#comment-916 Matt Tue, 21 Oct 2008 17:43:40 +0000 http://mad-computer-scientist.com/blog/?p=80#comment-916 @pcdinh: "Constraints and concurrency are not enforced by default: I love actually. By moving the logic of maintaining relationships into application developers’ responsibility, applications can be made more flexible and user-friendly." Until one of the app devs "forgets" to validate, or does not understand the data relationships, resulting in inconsistent data. What happens to user-friendliness then? Fail. A strength of a properly designed _relational_ database is the management of relationships in a single location, without exception. Use the right tool for the task. Secondarily, unfriendly error messages are a failure of application developers who do not know how to properly detect or catch exceptions and handle them properly. @pcdinh: “Constraints and concurrency are not enforced by default: I love actually. By moving the logic of maintaining relationships into application developers’ responsibility, applications can be made more flexible and user-friendly.”

Until one of the app devs “forgets” to validate, or does not understand the data relationships, resulting in inconsistent data. What happens to user-friendliness then? Fail.

A strength of a properly designed _relational_ database is the management of relationships in a single location, without exception. Use the right tool for the task.

Secondarily, unfriendly error messages are a failure of application developers who do not know how to properly detect or catch exceptions and handle them properly.

]]> By: pcdinh http://www.mad-computer-scientist.com/blog/2008/10/18/what-i-hate-about-mysql/comment-page-1/#comment-918 pcdinh Sun, 19 Oct 2008 15:19:54 +0000 http://mad-computer-scientist.com/blog/?p=80#comment-918 The lack of full support to full outer join is not a pain http://www.xaprb.com/blog/2006/05/26/how-to-write-full-outer-join-in-mysql/ Constraints and concurrency are not enforced by default: I love actually. By moving the logic of maintaining relationships into application developers' responsibility, applications can be made more flexible and user-friendly. At least error message will be more friendly. Stored procedures: It is overrated. SP can be a maintenance nightmare, especially when you do heavy refactoring The lack of full support to full outer join is not a pain

http://www.xaprb.com/blog/2006/05/26/how-to-write-full-outer-join-in-mysql/

Constraints and concurrency are not enforced by default: I love actually. By moving the logic of maintaining relationships into application developers’ responsibility, applications can be made more flexible and user-friendly. At least error message will be more friendly.

Stored procedures: It is overrated. SP can be a maintenance nightmare, especially when you do heavy refactoring

]]> By: Mariusz http://www.mad-computer-scientist.com/blog/2008/10/18/what-i-hate-about-mysql/comment-page-1/#comment-917 Mariusz Sun, 19 Oct 2008 15:19:29 +0000 http://mad-computer-scientist.com/blog/?p=80#comment-917 What can I say, I agree with you... MySQL with default MyISAM engine is only pseudo relational database. There are a lot of other limitations in MyISAM, e.g. transactions or table locks. PostgreSQL is much more powerfull open-source database system. If I can, I always choose this RDBMS. Maybe PostgreSQL is a little bit slower for smaller databases, but for medium or large... What can I say, I agree with you… MySQL with default MyISAM engine is only pseudo relational database. There are a lot of other limitations in MyISAM, e.g. transactions or table locks.

PostgreSQL is much more powerfull open-source database system. If I can, I always choose this RDBMS. Maybe PostgreSQL is a little bit slower for smaller databases, but for medium or large…

]]> By: Thomas Edwin Santosa http://www.mad-computer-scientist.com/blog/2008/10/18/what-i-hate-about-mysql/comment-page-1/#comment-914 Thomas Edwin Santosa Sun, 19 Oct 2008 10:37:05 +0000 http://mad-computer-scientist.com/blog/?p=80#comment-914 Fourth: no easy way for raising exception Fifth: no easy logging Fourth: no easy way for raising exception
Fifth: no easy logging

]]> By: Jonathan Ellis http://www.mad-computer-scientist.com/blog/2008/10/18/what-i-hate-about-mysql/comment-page-1/#comment-913 Jonathan Ellis Sat, 18 Oct 2008 17:16:18 +0000 http://mad-computer-scientist.com/blog/?p=80#comment-913 Last I checked, even innodb didn't honor CHECK constraints. Last I checked, even innodb didn’t honor CHECK constraints.

]]> By: Liam http://www.mad-computer-scientist.com/blog/2008/10/18/what-i-hate-about-mysql/comment-page-1/#comment-912 Liam Sat, 18 Oct 2008 16:37:52 +0000 http://mad-computer-scientist.com/blog/?p=80#comment-912 I got one. If you have a table with two TIMESTAMP columns, and update one of them, both will be updated. This is because MySQL didn't have triggers for a very long time, and thus created a data type that updates every time you change a row. I got one. If you have a table with two TIMESTAMP columns, and update one of them, both will be updated. This is because MySQL didn’t have triggers for a very long time, and thus created a data type that updates every time you change a row.

]]> By: marco http://www.mad-computer-scientist.com/blog/2008/10/18/what-i-hate-about-mysql/comment-page-1/#comment-911 marco Sat, 18 Oct 2008 16:23:10 +0000 http://mad-computer-scientist.com/blog/?p=80#comment-911 and the fact that tables with ibdata aren't shrinkable after a delete ?? and the fact that tables with ibdata aren’t shrinkable after a delete ??

]]>