I know very LITTLE about the internal working of MySQL. So, I am not the right people to comment it. However, from my experience MySQL is quite optimized for some operation but not for all.
I think Julian actual suggested a very good idea on the searching optimization for single user. At present, I don't have time to have a serious examination on all their implications (esp. on multi-user cases which is the actual working environment of MySQL). May Julian tell me more about your suggestions' implications on multi-user environment? For example, 2 or more processes updating the same table at the same time. Best regards, KH > On 23-Feb-2003 Julian wrote: > > Speed improvement with packet proceeding!? > > > > 1. Packet proceeding: > > > > I'm not quite sure is it possible with Mysql but it could be "easy" to > > be implemented. > > For example: > > > > select * from table where id=52 and name='some' > > select * from table where id=23 and email='[EMAIL PROTECTED]' > > > > these two queries select row(s) from table 'table' which means that > > these two queries could be tested simultaneously, so database file will > > be proceed only once. > > What about the case where one (or both) selects fail? > > And how would the application tell that there were multiple > rows where id=23 and email='[EMAIL PROTECTED]' ? > > If you *know* that these two rows exist and unique then: > > SELECT a.*,b.* from table as a, table as b > WHERE a.id=52 and a.name='some' AND > b.id=23 and b.email='[EMAIL PROTECTED]' > > would do the same thing. > > <snipage> > > > > > In fact I'm not talking only about 'select'-s but any tables > > examination/traverse (i.e where clauses etc..), so queries like these > > could be also speed up: > > > > update table set data='test' where name='some' > > select * from table where id=10 > > > > What would be the sensible error message if your update failed > but your select succeeds ? > > What would be the expected value of mysql_numrows() ? > > And what if there are multiple rows where id=10 ? > > <snipage> > > > 2. Bulk update/delete etc.. > > > > Take a look at this update query (not implemented.. yet!) > > update table1 set column=value,... where clause limit #, update table2 > > set .... > > (or delete from table1 where clause limit rows, delete from table2 where > > clause...) > > > > The same question: How would a program tell which statement failed? > > <snip again> > > > Have I a good point here? Any ideas and discussion about these > > suggestions? > > You'll need to rethink the failure modes. > What you're suggesting might be useful in certain special case(s) > but badly fails the general case. > > Regards, > -- > Don Read [EMAIL PROTECTED] > -- It's always darkest before the dawn. So if you are going to > steal the neighbor's newspaper, that's the time to do it. > (53kr33t w0rdz: sql table query) > > --------------------------------------------------------------------- > Before posting, please check: > http://www.mysql.com/manual.php (the manual) > http://lists.mysql.com/ (the list archive) > > To request this thread, e-mail <[EMAIL PROTECTED]> > To unsubscribe, e-mail <mysql-unsubscribe- [EMAIL PROTECTED]> > Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php -- Yours, KH Chiu C&A Computer Consultants Ltd. Tel: 3104 2070 Fax: 3010 0896 Email: [EMAIL PROTECTED] Website: www.caconsultant.com --------------------------------------------------------------------- Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
