18 messages in com.mysql.lists.bugsRe: [software] Re: Update problem
FromSent OnAttachments
Alexander Aulbach04 Dec 2000 15:46 
Sinisa Milivojevic05 Dec 2000 05:28 
Alexander Aulbach06 Dec 2000 04:39 
Sinisa Milivojevic06 Dec 2000 04:55 
Alexander Aulbach06 Dec 2000 06:58 
Michael Widenius06 Dec 2000 10:56 
Sasha Pachev06 Dec 2000 11:39 
Alexander Aulbach07 Dec 2000 06:42 
Michael Widenius07 Dec 2000 17:39 
Alexander Aulbach08 Dec 2000 04:22 
Michael Widenius09 Dec 2000 04:38 
Werner Maier09 Dec 2000 06:51 
Michael Widenius10 Dec 2000 07:23 
Alexander Aulbach11 Dec 2000 07:10 
Alexander Aulbach13 Dec 2000 07:02.tgz
Michael Widenius14 Dec 2000 23:04 
Alexander Aulbach15 Dec 2000 03:28 
Michael Widenius15 Dec 2000 16:51 
Subject:Re: [software] Re: Update problem
From:Alexander Aulbach (ssi@mainmedia.de)
Date:12/08/2000 04:22:53 AM
List:com.mysql.lists.bugs

On Fri, 8 Dec 2000, Michael Widenius wrote:

Could you please ftp the binary data for the above 'texts' file (not a dump) to ftp://support.mysql.com/pub/mysql/secret and I will take a look at this!

I just put the hole texts.* files to that URL

Oh, no not exactly. I first put it to /pub/mysql/Incoming and then to pub/mysql/Incoming/secret and then I read your mail once more and put it /pub/mysql/secret but I couldn´t overwrite the files, so I assume it is a symlink. :-)

PS: Anyway, note that it will always be faster to not include the key columns in the SET PART as in this case MySQL need to do a lot more work to find and update the row.

Yea, but the query was generated by automatic with the query_sql.inc I have written for PHP-Lib. It´s very complicated to decide wether it is a key and if it is a key, wheather it must be updated - in the wirst case I need one SELECT statement for every key, to decide this, so I thought it should be faster to just update it. :)

(You can't simple do a read-update-read_next loop when you are modifying indexes that you are using for retrieving data)