| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Comments:
<0> but ive started learning some Ruby on Rails stuff and they do a bunch of that
<1> that's what I don't get: the linking is done in the queries, not the tables..
<0> show me an example query your talking about
<1> anything left join really
<0> ?
<1> select foo from bar left join blah on foo.bleh = bar.blub
<1> *blah.blub, of course..;)
<0> you got me
<1> gotcha..;)
<1> anyway, I'll keep that stuff turned off, how is it gonna create pdf without the lib anyway..
<2> archivist: available now ? :P
<3> yup
<4> hey, I was going to ask you guys, Apachez & archivist, if you can just separate your word table into different partitions
<4> instead of a bunch of different tables
<4> or maybe there's a limit on partitions
<3> probably, google does
<4> so why not just do that instead of a new table for each prefix?
<2> how you mean ?
<2> ahh just using plain partitions ?
<2> na in this case well ehh
<2> let me just have the prefix method instead :)
<2> the fine part with prefix is that it scales over multiple servers so i know that server1 takes care of prefix aa->da for example
<5> Hi! (again) lol
<2> Jivedue i guess you have seen my output on archivists site ?
<2> the explain select and all that
<4> some of it, I think
<2> http://www.archivist.info/search/index.php/Talk:Sql
<2> http://forums.mysql.com/read.php?24,74774,74774#msg-74774
<4> I've just been following some your guys' conversations, because I find the subject interesting
<2> so far i have found that the method of one select per search word is the fastest
<2> but i guess there should exist some sort of inner join or something like the idea archivist came up with which should be faster than running 10 queries one after another
<6> Hi!
<6> Why does MySQL always build a 'test' DB upon installing? (I know, I should go to #mysql)
<3> why small words are needed: using almost ANY search engine find the mysql in() function, you cant
<7> hey anyone know where are the mysql logs in rhel
<7> servers
<2> Nilesh: check the my.cnf file
<8> sgershon: I bet it has to do with testing. :)
<6> mendel, LOL.
<2> archivist: yeah in my case i have set a lower limit of 2 chars and higher limit of 32 chars
<7> i am using 4.0.22
<8> Also, you are in #mysql
<8> So I am confused
<2> SGershon: the test db is used to test connections against when you compile them... like with DBI/DBD in perl etc
<6> mendel, it actually already build the 'internal' mysql database for holding tables and users...
<2> the mysql db is for logininfo and stuff
<7> mysql logs?????
<2> the test db is to test stuff with
<7> i need path
<2> Nilesh: as i said, check your my.cnf
<2> find / -name "my.cnf"
<7> its not there in my.cnf
<6> Apachez, I see. So these packages use the "test" table automatically, and if it is missing, they may present problems?
<3> I allow any size including numbers (A,B,C etc is a valid name or initial in my data)
<2> if the test db is missing you will get errors that tests failed
<8> sgershon: For example, DBD::mysql will connect to MySQL to test itself.
<2> the tests usually test stuff like create table, delete table etc
<2> to make sure it is working properly
<6> Sounds logical. Thanks Apachez.
<7> [safe_mysqld]
<7> err-log=/var/log/mysqld.log
<7> pid-file=/var/lib/mysql/mysql.pid
<7> open_files_limit=8192
<7> found it thanks
<2> what i dont get it is why the query time is getting that high when i tried the archivists method
<4> Apachez: are you wanting to allow search operators and quotes, etc.?
<3> the optimiser may have used the word with the highest qty first which is why it needs rewriting as a straight join
<7> how to upgrade locate commanddata
<7> "locate"
<4> or do you just ***ume all words must be present?
<2> quotes would be nice but basically all i need is to search for each word (in AND mode) and to allow wildcards
<7> its not locating directry which is already present
<4> k
<5> well guys... back to the same issue...
<5> well... this is the case: I have one field in a table with Number and Street, like '123 Patafurdio ave.' some cases has only the street name... i want to copy the Nrs from this field to a new filed, and Some of them are like '123B', what's the quickes way to copy them to the new new filed?
<7> updatedb?????
<5> field*
<7> i forgot command?
<3> I dont allow wildcards but do think a thesaurus could be usefull
<5> this numbers are all separated from the street name by a space...
<2> archivist: so basically no suggestions for my methid i pasted on your page then ?
<7> Apachez????
<7> updateDB command???
<2> Nilesh ?
<5> this database has 400.000 records... should i do this all by hand?? (LOL)
<7> i mean what is the command to update locate databse?
<3> mainly cut down on the qry eg move the limit to earlier in the search (which separate qrys will allow) cut down/restrict the wildcards if pos, all to reduce the result qty's before the limit 500
<7> Apachez: what is the command to update locate databse?
<2> Nilesh: how should i know? i use the find command
<7> hehe
<7> its ok
<9> updatedb
<3> -U or wait till cron does it
<10> yeah its updatedb
<10> >> /usr/bin/updatedb
<2> archivist the problem by cutting down is that there will be inaccurate results
<2> like if i limit to 10.000 in the mulitple qrys then there is a chance that i would need 11.000 as limit to get all 500 rows which has all search words
<3> the results are inaccurate as you through excess by the limit
<2> nope
<7> yeah thats it LinuxGocha
<3> through=throw
<2> the result is accurate since its in the last scan it will limit to 500
<2> like example.. if you use limit 10000 for the small queries that search for each word then the result might be
<2> word1: rowid 0-9999 word2: rowid 10000-19999
<2> which means that 0 rows will get returned to client
<2> compared to if limit would be 20000 rows in the first scan giving word1: rowid 0-19999 word2: 10000-39999 which gives that rowid 10000-19999 has both searchwords
<2> and in the end there is a sort by date and the top 500 rows gets returned to client
<3> google for the paper that the google founders wrote, its interesting
<2> any ideas what the name might be ? :P
<3> must be 4 to 6 yrs since i downloaded it
<10> >> /scripts/restartsrv_cppop
<10> help you Nilesh
<7> did
<7> thnx
<10> have you checked your forums ?
<10> new post are there :)
<7> which forum?
<10> http://tuxterminal/forums/
<7> ahh its my forum man
<7> do u post it?
<10> some of
<10> actully want to make it more good
<7> are that linux user if am not wrong
<3> Apachez http://labs.google.com/papers.html a start!
<10> yeah sir :)
<7> good posts
<7> http://tuxterminal/forums/
<11> hi all
<11> where can i find documentation for client version dependencies
<11> i.e. i want to upgrade my database server to 5.x but need to make sure that my 4.x clients will still work
<2> archivist: thanks but that doesnt actually help in this case :P
<12> hi, i've got a DB in dreamhost with latin_spanish collation, how can i change it to UTF8?
<9> mxm***ter: 4.what?
<11> 4.0.x and some 4.1.x
<9> well for the 4.0s it'll be a little harder
<9> !m mxm***ter upgrading from 4.0 to 4.1
<13> mxm***ter: (Upgrading from Version 4.0 to 4.1) : http://dev.mysql.com/doc/mysql/en/Upgrading-from-4.0.html
<9> !m mxm***ter upgrading from 4.1 to 5.0
<13> mxm***ter: (Upgrading from Version 4.1 to 5.0) : http://dev.mysql.com/doc/mysql/en/Upgrading-from-4.1.html
<12> hi, i've got a DB in dreamhost with latin_spanish collation, how can i change it to UTF8?
<9> hi we heard you the first time. if someone has an answer they'll let you know
<10> freemon i think it will only change when your hosting provider will change it for all server
<11> shabbs: what that doesn't say (and what i am interested in) is will a 4.1.x client work with a 5.1 server?
<11> and will a 4.0.x client work with a 5.1 server?
<10> i thing it is a Default
<12> LinuxGocha, if i make a new table
<12> how can i make it in utf8 mode?
<9> mxm***ter: a 5.1 server? I'm not sure. I don't know anything about 5.1 as it's still beta. 4.1 will/should work with 5.0 servers. 4.0 clients will not work with 5.0 without an upgrade/tweak on server side
<14> hello!
<9> hola
Return to
#mysql
or
Go to some related
logs:
gentoo fatal inserting nvidia unkown symbol in module
#perl
regex if followed by a decimal
cannot encript in mysql
#perl
#css
#kde
php proxy code
#gentoo
#web
|
|