http://pastebin.ca/698448 Leith

does mysql not have FULL OUTER JOIN? dang this sucks

that is the my.cnf not the error log no

i think it does suck

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

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

Leith, i fixed it

works for me

nm had user agent set to googlebot, but thanks i’ll consider it Leith

how do i know if MySQL has the MyISAM engine included?

SHOW ENGINES;

MyISAM is *always* included and on

yea i figured that out :3 thanks what is mysql_extend_adoconnection Class pear_extender: Cannot inherit from undefined class mysql_extend_adoconnection in /homepages/44/d211289343/htdocs/fatchan/lib/adodb/generic_modules/pear_module.inc(5) : eval()’d code on line 1

“#1267 - Illegal mix of collations (utf8_unicode_ci,IMPLICIT) and (utf8_general_ci,IMPLICIT) for operation ‘match’ “

Class pear_extender: Cannot inherit from undefined class mysql_extend_adoconnection in /homepages/44/d211289343/htdocs/fatchan/lib/adodb/generic_modules/pear_module.inc(5) : eval()’d code on line 1 help :/ i dont know what this is http://code.google.com/p/kusaba/wiki/InstallationGuide

hi, I am using mysql C API and I’m trying to insert some data into a table; however I get an error and can’t figure out what is the problem: http://rafb.net/p/7ODw5N64.html Can anyone help?

Is there a way to make a field that will automatically index entries?

you mean an autoincrement column that is the primary key for a table? that’s rather standard

yeah I knew it could be done My memory is a bit, umm, dusty about 3 years though. So just create a primary key column?

just google up some lame tutorial, should be in any

is it recommenadble to use a primary key for just one column or all of my columns?

hi there

hello

i’m founding some problems installing mysql on OS X I want to install it to run wordpress on my local host What I did is downloading the dmg packege installing mysql and mysqlstartup now following the guide I run sudo /Library/StartupItems/MySQLCOM/MySQLCOM start usr/local/mysql/bin/mysql ERROR 2002 (HY000): Can’t connect to local MySQL server through socket ‘/tmp/mysql.sock’ (2) any idea?

look at the config file and try to change socket configuration

where is it, in /etc?

search for my.cnf. I’m not a mac user, so I can’t tell you

ok

Folks, you know how you can say like, grant all privileges on DBNAME.* to…. Is this “*” notation just supporting the star, or you can say something like DBNAME.??* like in shell globbing ?

I can’t find it ron_asheton, I looked in etc and usr/local/mysql-version…/ cd /usr/local/mysql sudo chown -R mysql data/ why do I should change the owner? can you understand?

no we can’t sei da italia, Matteo_M ?

si checking the /tmp folder I dont find it as the guide says: “The MySQL socket file is created as /tmp/mysql.sock by default”

which platform ?

OS X

Matteo_M, ln -s /var/mysql/mysql.sock /tmp/mysql.sock try that

ok seams to work I get a different error Access denied for user ‘matteo’@'localhost’ (using password: NO)

Matteo_M, mysql -u root

what is -u?

it tells mysql to enter as ‘root’ user. then you should set a password.

so it’s the same if I run it using the sudo command or not ron_asheton? sudo /usr/local/mysql/bin/mysql

no, mysql has its own set of users… sudo runs commands as the root user of your local computer

sudo -H mysql if you want to run it that way

ok thanks guys

what is the max for an INT please ?

can someone tell me the command for adding adding a range of rows where nid is 100-200 and tid is all 3?

http://dev.mysql.com/doc/refman/5.0/en/numeric-type-overview.html

d

Error on rename of ‘./ndb/ to ‘./ndb/test’

thanks olafski

hey is that compile time username-length patch in mysql 6 or not? usernames should be 128 chars like mysql identifiers.

Class pear_extender: Cannot inherit from undefined class mysql_extend_adoconnection in /homepages/44/d211289343/htdocs/fatchan/lib/adodb/generic_modules/pear_module.inc(5) : eval()’d code on line 1 help? :/ im trying to install Kusaba http://code.google.com/p/kusaba/wiki/InstallationGuide

how can I set the database default coalition from latin1_general_ci to utf8_general_ci

ALTER database yourDb COLLATE collationName

Jax, not just one database, the default for new databi

how hard is it to read csv into mysql as a table?

robboplus, I don’t know with mysql itssself, but with phpmyadmin, it’s easy

!man import

see http://dev.mysql.com/doc/refman/5.0/en/alter-table.html

aarcane i don’t like nor use phpmyadmin… just row mysql raw*

Where can I find an official reference and tutorial for mysql?

robboplus, then I don’t know the answer, sorry.

I prefer a free one

Soroush81 type !man !man select

see http://dev.mysql.com/doc/refman/5.0/en/select.html

there you got everything

hello

Soroush81, the manual on the website for mysql is a good reference.

I try to find them

would you guys consider mySQL to be much more effective for data access in servers than a flat file?

but I have never seen a man with my mysql

Soroush81, http://dev.mysql.com/doc/index.html

I am thinking of putting some data separately in a flat file as it doesn’t fit the format of a table very well

How different is mysql from other sqls in commands and structure they use like Postqsql

Agamemnus, so do it

PLSQL, MSSQL

Agamemnus why wouldn’t it fit?

Soroush81, it’s pretty similar usage wise, but when you’re referring to administration, or advanced features, that’s where they differ.

When should I prefer mysql rather other sqls?

Soroush81, when all you need is raw speeeeeeed, and not much else

it’s a very small piece of data, one byte for each column in one of my tables

aarcane you find your answer?

speed is nice.

was downstiars

Jax, nope Jax, I need to change the default for the whole server, not just the current for a database

yeah set it in my.cnf

wouldn’t make much sense to put this in a string in a separate table made of one column, would it?

collation-server=latin1_swedish_ci you can probably also just do SET collation_server=utf8_general_ci aarcane but will be set back to default when you restart mysqld, so put it in my.cnf archivist you around? Agamemnus what are you doing? sorry just saw what you were talking about Leith you here? i need to annoy somebody with an idea

Jax, when I put it in my mysqld section of my.cnf, is it SET or just collation_server = utf8_general_ci Jax, when I put it in my mysqld section of my.cnf, is it SET or just collation_server = utf8_general_ci ? oops, sorry for the double entry

you just put collation_server = utf8_general_ci in your [mysqld] section

email it to me! (wife’s birthday party atm)

ah sorry for bothering! party on!

thanks Jax

time to go home bye bye!

eee

Class pear_extender: Cannot inherit from undefined class mysql_extend_adoconnection in /homepages/44/d211289343/htdocs/fatchan/lib/adodb/generic_modules/pear_module.inc(5) : eval()’d code on line 1 trying to install kusaba, http://code.google.com/p/kusaba/wiki/InstallationGuide

Sounds like a ##php question

does anyone know what this could be?

hello people) can anyone tell me is there support for views in mysql 4?

There is not

too bad for me..

Just upgrade

mysq local serv has 5 version… the problem is the hosting provider of my client….

Switch providers

he won’t

They always say that

yeah

It is far more cost effective to have developers use the right tools, than to cripple them by forcing them to use out of date ones to save a tiny amount of cash

)))

How much does “he” pay per month for the hosting provider?

Also, as a general point, you should be using the same version of mysql in your development environment (as well as everything else)

dunno thanks anywa

evening are the collation formats somehow standarized in mysql? i.e is there a standard that defines utf8_general_ci or something? the format of that string i mean. in my oppinion, it’s neccessary to be able to derive collations from locales. since locale’s are standarized

select *,max(users) as usercount from bla where foo like “%bar%”; select *,sum(users) as usercount from bla where foo NOT like “%bar%”; . Is it possible to combine this in a single query?

mysql doesn’t have collations for every locale… and the collation really depends more on the data than on the client, in the end and then there are further complications, such as the existence of two separate collations for spanish.. i suppose the modern collation would correspond to certain locales, while some locales may still be using the traditional collation… pretty complex, though, and not something mysql would be keeping track of utf8_general_ci is a pretty bastardized thing in the first place, not really following any standard.. should use utf8_unicode_ci if you want to follow some standard

sounds like you really should be using fulltext indexes doing MATCH(foo) AGAINST(’+bar -bas’ IN BOOLEAN MODE) see: http://dev.mysql.com/doc/refman/5.0/en/fulltext-boolean.html

thanks for the hint

%bar% will *never* use an index so performance will be hell if you don’t use FT

well, thats not the main problem i wanted to describe, its that i want to select max(users) as usercount OR sum(users) as usercount in the same query, depending on the value of another column arr, that doesnt make sense its one query but with a lot of UNION ALL , nevermind ,i cant describe it correct right now , max(users) where foo=0 as usercount1, sum(users) where foo=1 as usercount2 from table;

I have the following query: UPDATE feed_lists, item_tags, user_tags SET feed_lists.tag_id = 12, item_tags.tag_id = 12, user_tags.tag_id = 12 WHERE feed_lists.tag_id IN (SELECT id FROM tags WHERE tags.name LIKE ‘Apple’) I need my conditional to match where feed_lists.tag_id IN ( [subquery] ) and for the other two tables How do I do that without another OR and doing 3 more sub queries 2 more*

indexed views in 5.0?

uhhhhh Is there anyway to revert a database from a query you may have made? Is there basically an undo query button?

rollback; if you’re running innodb and haven’t committed.

MySQL :-\

heh

I ran the query on my Live server

yes, of cours mysql. mysql has isam and innodb engines.

innodb is not a database server

arggg…. I use mediatemple…..

so INDEXED VIEWS anybody know?

I don’t think they use….innodb…

Not ISAM anymore

*thinks he just owned himself

what is it?

i found a thread on it at the mysql forum from 2005, where a MySQL develop said AB wanted to add it soon

I usually assume if they can ROLLBACK, they know they’re in a transaction MyISAM. ISAM is the old version taken out in recent versions.

meh, close enough.

Hehe, well, there *was* an ISAM that some people still use.

your only real hope is to restore from a backup and use the binary log to roll forward, if you have it enabled

binary….log?

nice. !man binary log

see http://dev.mysql.com/doc/refman/5.0/en/binary-log.html

well it’s okay, it’s only 3 tables I guess I can…..roll back those 3 tables they weren’t a huge deal Thank god for backupos backups* Well, last time I test anything on a live server.

rofl @ “it’s only 3 tables”

lol…why?

what do you mean, why? 3, hell 1 table could be huge

Well they are pretty big, I altered one very important column :’( well actually they aren’t really big lol It’s impossible to fix manually MySQL needs a temporary roll back feature wow that’s surprising phpMyAdmin or MysQL doesn’t have a way

um, it has rollback. don’t you listen?

haha

????? you mean innodb?

You are more than welcomed to program that functionality into MySQL

how can mysql not have indexed views yet? i dont’ understand this fuck fuck fuck fuck ass ass shit bitch aksdgj asg

Let’s keep the language clean

why haven’t you written the code for it yet?

I think my host uses: MyISAM So am I shit out of luck?

because indexed views is not the optimal solution for my problem, only an acceptable one. if i end up writing the code myself, it won’t be indexed views for mysql

you can create your tables with either engine.

man I have no idea what to do, I guess I’ll just rollback my tables…they haven’t gotten much activity. I just need the easy solution I guess And I don’t know how to rollback by using MyISAM

You can’t

Great!

You really, really need to test major changes on a test server

~9200 INSERT’s (ID and 26 VARCHAR(45)’s ) in 300 seconds on a Intel Core Duo 1.66 GHz… I was hoping for more

if your application is to be supportable, you’d better have a rollback plan for each change too Was that individually, or as a single batch? What table handler? WHat other columns existed? how many indexes?

^ The stats is for MySQL5 and a loop in PHP.

you can’t roll back DDL anything, this is a pointless discussion

MarkR42 - Individual. InnoDB

You can get a lot more performance by doing inserts as a batch, especially with innodb. Did you have autocommit enabled?

well this should be fun to download 180 MB backup! At least I am downloading at 1 MB/sec

MarkR42 - Don’t now about autocommit. But I consider writing a temporary file and insert data using LOAD command instead.

2 minutes! yay Man what did they do back in the dial up days…sheesh

Whenever we do a major change, we always have a roll back plan - in some cases, this would mean renaming a table and creating a new one with the new schema, and using insert … select Keep your backups close to where you need them

kalmar, create temporary table loader(blah ..) engine=memory; .. when done, insert ignore … select … from loader; .. drop temporary table loader;

If you are doing 9k inserts individually, probably turnaround and context switch time is signficiant.. you also want to watch the IO as innodb is probably doing a flush after every single one if autocommit is enabled Try doing large batches using extended insert syntax, or failing that, a LOAD DATA

then you save yourself trouble when someone fucks up the data, but you miss out on the real purpose of backups: disaster strikes your HDs

So, I have a huge database when many tables, anyone know a way of me being able to pick on the tables and their data that I Want?

… what?

I downloaded by backup, how do I pick out the queries for certain tables?

Thanks for the suggestions. I might use LOAD DATA as the last option, not that I don’t like (because I really do!) but it’s always fun to learn something new. Like experimenting with autocommit

opening it in notepad and then doing it seems a bit rough doesn’t it?

your backup is a list of queries? how much do you get paid?

Most of our data are not backed up, for completely logistical reasons The important stuff is

Logistical reasons?

i was only trying to point out a downside to your advice of “keep backups where yuo need them” the data is probably updated frequently / becomes outdated frequently

there isn’t physically enough space to back up all of our data, nor would our servers have enough time to back it up if there was.

donaldo, it’s a back up that MySQL would normally give you or any GUI tool.

or it contains incriminating evidence and they don’t want any backups to help convict them

heh

The data do change frequently, and they are very large

Yes, my back is a bunch of queries that contain the data I need to import. This isn’t something abnormal, and I don’t get paid at all.

We work on the basis that the failure of a single machine will lose us sufficiently few data that we won’t mind too much Although all the discs are RAID, some types of failure could trash a complete box

It’s very common for your MySQL dumps to be like this, phpMyAdmin gives you the same dump

huh, strange. i’ve only ever backed up databases by making copies of the underlying files stoerd by the database server… of course this is a very wasteful process for most applications

Well, you don’t develop many websites do you?

Making copies of the files is a very GOOD way of backing up myisam, in fact, probably the best

you could use sed or grep to strip down the backup to only the queries that are relevant really? why?

Unfortunately, to back up myisam consistently you NEED to do a “flush tables with read lock”

I use Windows.

i hope YOU don’t develop many websites

Because copying the .MYI and .MYD is much faster than mysqldump

donaldo, why is that?

so if you do a “flush tables with read lock” *anyway*, you may as well copy the files, provided you use myisam exclusively

i don’t care if you use mac os 9. if you don’t have a tool to run a regular-expression on text files, move out of your computer cave and get some real fuckin tools

I’ve never needed it buddy.

hmm… wouldn’t a real filesystem backup tool be better though? on many filesystems, it could detect exactly which blocks on the HD have been modified, thus making it a faster process

lol

ok “buddy” have fun using Notepad then

Possibly, I’ve never investigated that. I’ve heard about people using lvm to backup

eh, i’d use hardware raid solutions first in most situations if i wanted to go that route

there is some kind of lvm clone option, which creates a copy-on-write clone of a filesystem, which you can then backup while mysql continues to write to the original

haha, it’s okay, you’re probably 45 and badass because you know MySQL

don’t get me wrong, i love LVM, but if i’m doing a serious job, no reason to slow down the system with that extra work

We don’t use it in production

I understand you’re bald and bitter.

dude, i don’t know jack squat about mysql, honestly. and i’m 22.

Let’s keep personal attacks to yourself?

wow lets see how long this unprofessional stuff lasts in this professional channel.

Not long!

Good, then shut the fuck up

Who said this was a professional channel anyways?

heh couldn’t you tell how professional it was from the formal and courteous interaction between myself and Delixe?

there are professional irc channels? thats news to me!

No, not really. But I don’t really pay too much attention.

seekwill, you may be right but if I want a Jerry Springer Special (Delixe) .. I’d tune in whenever it’s on

Still playing DTD??? hehe

not right this moment… but I might start again now that you mention it…

MarkR42 Tapout - Thanks guys! Speed is fine now! I’ll stick to SET AUTOCOMMIT=0; [Bunch of queries]; COMMIT for now, at least in this application. It also gives me the opportunity of using ROLLBACK which I didn’t think of at all.

If it was a Jerry Springer Special, we’d be cousins, I’d be having sex with his mother outside in the back trailer everyday and donaldo would really be a woman also having sex with his mom.

C’mon, let’s keep that out of here?

Tapout brought it up.

I don’t care.

me neither.

You may find that varying the number of inserts between COMMITs is helpful also, doing extended multiple inserts is a really good idea

never heard of extented multiple inserts…

Look up insert, insert into tbl (a,b,c) VALUES (1,2,3), (4,5,6) etc you can potentially do a whole bunch of inserts in a single query, this improves performance a lot especially on a network, where no more round-trips are required standard Mysql only allows a single INSERT per query

logical syntax and a must try

the nifty bit is, the table handlers do optimisations on this, like writing all the key updates in one go e.g. on innodb, I think it does it all in one txn even if autocommit is on reduces IO a lot *potentially*

Lovely! $res = mysql_query($query); oops

Of course the number of rows you can put in depends on the max packet

load data infile

but still, that’s usually rather a lot (depending on the row length). Yes, “load data infile” is *probably* better - at least, so I’ve heard. I’ve not benchmarked it for our app, it’s on our “to do” list

lol LDI is WAAAAAAAAAAAAAAAY faster. It may not matter for 10 rows…

If you’re on a network, LOAD DATA LOCAL INFILE transfers the file, right? Basically, none of our inserts happen locally, everything is networked

More than once, where a user was inserting from an INSERT query, it takes…10+ minutes. LDI makes it less than 30 seconds. Correct LOCAL indicates the file is local to the client.

I used LOAD DATA INFILE in an application where I extracted text from pdf’s and made them searcable. I works extremely well!

the mysql client library must read the file locally and send it to the server somehow, over its socket, but it’s cool that it does all that for you efficiently.

I am having a problem of with left join. I get extra results with NULL values at the value I extract from the table I am left join with. this is a pretty minimal query that reproduce the problem: SELECT h.url_id, meta.value FROM `wp_22_firestats_hits` h LEFT JOIN `wp_22_firestats_url_metadata` meta ON h.url_id = meta.url_id; any idea what can cause this? damn, I got it. forgot to filter other fields from the metadata table (where meta.type = ‘1′).

that’s how outer joins work

Flabobo, ?

skip the left word and see if it does what you want

flupps|NO, ? flupps|NO, no, without the left join it removes records that are not in the meta table. and this is not what I want. specifically, the metadata table contain url titles. I want to extract the url ids and the url title, or null if there is no title.

hye buddies

I found my mistake.

your query above should work, then

whats wrong with this http://rafb.net/p/TpG2Cy99.html SQL Syntax error…

[ and ]

flupps|NO, the problem was that the metadata table contain other things as well, not just title. and I needed to filter the other things out.

flupps|NO , so , how do i create a table only if it doesnt already exists ?

create table if not exists (and so on…) brackets indicates [WHATEVER OPTION] is optional *indicate

oh lol thanks

yw

[root@game ~]# /etc/init.d/mysql status ERROR! MySQL 5 is not running, but lock exists

flupps|NO, hmm, actually my solution is not good. it filters results that does not have the metadata of type 1, while I want those results to appear with null value.

Put it in your join condition I’d open that start up script, and look for the lock file it looks for, and delete it…

seekwill, hmm, that seems to work. thanks.

np

I have a sense of humor, but how do I fix a messed up database?

Hi! This might not be a mysql related question, but it might as well. I’ve set up syslog-ng to log to mysql. It writes a query to a FIFO pipe, and then I’ve got a bash script running which pretty mich does while [ -e /tmp/mysql.pipe ] myql -bla .. /tmp/mysql.pipe.. Now the thing is that each log entry is inserted three times Is this something to do with the fifo and mysql, or is it actually written to the pipe three times (does anyone know how to check this?)

http://pastebin.ca/698847 — how do I fix that?

!perror 145

Table was marked as crashed and should be repaired

!man repair table

see http://dev.mysql.com/doc/refman/5.0/en/repair-table.html

But in general, you fix it by not having crashed tables; always shut down mysql properly if possible. This is especially important if your table is several Gb and has loads of indexes

the proper way being /etc/init.d/mysql stop?

when REPAIR TABLE will take a long time The proper way is to shut it down mysqladmin shutdown, init.d stop script etc And try not to have it crash

It shouldn’t crash at all… Any crash should be considered a bug to be reported.

If there are broken tables, there is probably a reason for it, you should really find out what the reason is and try to stop it happening again

server wedged and needed a reboot

Crashes can happen for reasons other than bugs, f.e. you m ay h ave dodgy hardware In which case, the reason for the server being “wedged” needs to be investigated so it can be avoided in future

Possibly. I’ve seen many more crashes due to bugs and being linked to incompatible libraries than hardware errors. ECC!

REPAIR TABLE mourd”?

It will take hours if the table is several gigs

table is relatively small

Is myisamchk faster?

However, you must do it as the table will be inaccessible currently I don’t think so

” on it

;

ah, ok

same as any other command in mysql CLI

I’ve never touched the mysql cli before hmm, no database selected… I guess I gotta do that first. /me googles.

seekwill, where would i find the lock file in the startup script

Look for it

well i thought i found it but the only reference refers to a location that doenst exist

awesome, fixed it. Thanks guys!!

i converted my tables to innodb last night, and now select queries are just… hanging. (

nm i got it i think [root@game subsys]# service mysql start Starting MySQL../etc/init.d/mysql: line 159: kill: (1963) - No such process ERROR!

Not sure how “start” invokes a kill..

how do you export just a part of a table that matches some query? to .sql that is basically i would like the data to be printed out to some .sql file instead of outputing them to the console in mysql shell is that doable?

sup

use a insert…select to load it into another table, then mysqldump it Unless that’s logistically difficult indeed it is How did you guess?

You’ve said it twice!

wait wait damn that idea sounds bright! seekwill hm nope, it’s beer or hm

You’ve said “wait” twice

crappy keyboard, that’s all ok i will be very honest with you.. never tried insert/select combo..

!man insert select

see http://dev.mysql.com/doc/refman/5.0/en/insert-select.html

oh he he thank you, reading up so what you mean is to just create a temporary table

Bingo You could try CREATE TABLE … SELECT as well

instead of a DELETE which will delete a large proportion of a table, create a new temp table, do an INSERT…SELECT to keep the rows you want, then drop the old table and rename the new one

“easier”

hm that’s a real brain fucker

say you have a 2G table which you want to delete 99% of the rows of, that is a good idea, otherwise you’ll end up with huge files emptyish of rows, making selects slower and using disc space

Put it on the wiki!

Will do

very nice stuff

OPTIMIZE TABLE

just have to think it all over and implement at once

optimize table sucks because it locks the table out completely while you’re doing it writes *and* reads a better alternative we’ve found is ALTER TABLE Engine=MyISA which only locks writes

You have benchmarks on that? How long that would take after the 2GB - 99%

Not yet, but we believe informally that in our particular case (loads of indexes) Bah, the wiki is telling me “could not open socket” when I try to register

how do I specify that I want an INNOdb table type when I create a database?

!man create table ok, that only works for seekwill

seekwill?

ENGINE=InnoDB

!man create table

see http://dev.mysql.com/doc/refman/5.0/en/create-table.html

!m hacim create table

hacim see http://dev.mysql.com/doc/refman/5.0/en/create-table.html

so: CREATE DATABASE blah ENGINE=InnoDB ?

WORKS FOR YOU TOO

can I use it at the DB creation level?

!m alter database

alter see http://dev.mysql.com/doc/refman/5.0/en/alter-database.html

no, I am creating a new DB

Ok, alter database does NOT have an option for default engine, only charset

not altering an existing one

no, just put ENGINE=InnoDB in each table in your create script shouldn’t be too tricky

ok

Can anyone tell me what might be wrong with this query: UPDATE feed_lists, item_tags, user_tags SET feed_lists.tag_id = 12, item_tags.tag_id = 12, user_tags.tag_id = 12 WHERE feed_lists.tag_id OR item_tags.tag_id OR user_tags.tag_id IN (SELECT id FROM tags WHERE tags.name LIKE ‘Apple’) I only want rows that match a certain tag_id to set with the new tag_id but instead the condition is always true… How can I do this without repeating the subquery for each of the OR conditionals?

It seems like you should be running that UPDATE for each table individually

okay. well…I didn’t want to because I am try to fix something and it would be tedious. Basically I’d have to do that same query about 30-50 times so times that by 3 and it becomes tedious

Right, but you only need three updates to update three tables

Well, I am trying to do some cleanup, maybe I’ll just write a PHP Script instead heh Basically, I Was trying to combine the 3 updates so I can just do it manually 50 or so times. If I split each one up I have to do it 150 times. Is there no way to updates multiple tables columns based on a sing conditional? single*

I have a process that doesn’t seem to want to die, and is locking my tables. I did a kill on the process ID but its still running

hm..

its been running for 2045 seconds, and is in state Killed i cant even stop the entire server unless I issue a kill -9 on the server PID or something

what’s the safe way to get rid of mysqld-bin.00000x files in /var/run? they’re occupying gigs and gigs of my hd space

If your server is not a replication master and you don’t want to do point in time recovery, use reset master BUT be ultra careful

reset master? and what am I being careful for the pc its running on is just a local dev box of mine not tied with anything else

in which case, “RESET MASTER” it is

yay it worked is there any way I can avoid those files from being made?

 

Affordable Phpnuke Web Hosting Solution

 

!m rage_ binary log

rage_ see http://dev.mysql.com/doc/refman/5.0/en/binary-log.html

hi guys, I need some help with counting rows. I have a table where I first read a portion of the rows (with a filter on, but limiting the amount of returned rows) and then do the same query again on the full table (with the filter, but without limiting), getting the amount of affected rows. What is the best way of getting this last information, the amount of rows that would be returned by the query without the prior limitation?

SQL_COUNT_ROW or somethign like that !m Hackwar2 SELECT

Hackwar2 see http://dev.mysql.com/doc/refman/5.0/en/select.html

anyone know how to solve this? [root@game subsys]# service mysql start Starting MySQL../etc/init.d/mysql: line 159: kill: (2061) - No such process ERROR!

Doing the same query again without the limitation and then counting the returned rows in php, doing the same query without the Select a.* part and then counting the returned stuff, doing the query with a count(a.id) or something else?

ive been using activerecord so much recently that my brain can longer recall mysql, can anyone tell me how i can delete a record if a field (a text field) contains a given string?

WHERE field LIKE ‘%string%’

thanks how do u write multiple conditions for that? WHERE field LIKE ‘%string%’ or WHERE field LIKE ‘%string%’ ?

Drop second where

and enclose it with brackets.

no

looks nicer…

Not necessary at all If that’s what floats your boat

no, but again, it _looks_ nicer

is it possible to regex matching on the LIKE condition?

Using less characters than necessary looks nicer. I think RLIKE might be it

sounds familiar

There is also a REGEXP function

at last..

does anyone have any idea about this? [root@game init.d]# service mysql start Starting MySQL../etc/init.d/mysql: line 159: kill: (3217) - No such process ERROR!

canadiancow, try service mysqld start

unrecognized service it worked yesterday so i dont know what happened i havent touched anything i rebotted the box thats about it

did you look up the /etc/init.d dir ?

huh? thats where the start script’s located

but is it still there ?

yes and i just tried replacing it with the one from the mysql source again to no avail

2 tables on tableA with primary key address, another Table with street_address and user_email primary key is user_id. There may be multiple users with tableB.user_email and a tableB.user_street_address which match tableA.address I want (but may not get ) a Select/join that will give me the equivalent of SELECT address, CONCAT((select tableB.email from tableB where `user_street_addr`=`address`)

address a primary key ?!??

Canadiancow. Try /etc/init.d/mysqld stop;while [ `ps -aef|grep -v grep|grep -c mysqld` -gt 0 ];do echo .;sleep 1;done;/etc/init.d/mysqld start it sounds likje it did not shutdown right and remove the mysqld.pid file so on startup it’s trying to kill it

ERROR! MySQL manager or server PID file could not be found!

Yes, oddly enough. It is for an HOA with a set list off addresses which is unlikely to change, unlike owner names, phones etc… so Using the address of the lot as the primary key. ps -aef|grep -v grep|grep -c mysql What does that return?

[root@game init.d]# ps -aef|grep -v grep|grep -c mysql 0

Hi all

Linux? if so which dist? MySQL ver?

I’m on linux, and i want to know, how can i reduce the number of the sleeping MySQL 5 processes i don’t know it’s name, i don’t know what i need to find… so that’s why i asks.

CentOS 5 x64

jgreene_greenemy, just a bit perplex… one same address can be written in multiple ways

mysql 5.0.45 but keep in mind, this worked yesterday

true, but the list is set, the addresses preentered and I’m working on the alg to clean addresses, mostly managing the user_street_addr at input time on the web interface end so it will match something in the master addr list for the HOA (which is already populated and won’t change). give me a minute. Need to check something.

k np

ever edit a n init script?

no ive never had the need to

ok.. one min then…

jgreene_greenemy, not sure to understand your request but you try to get people with no matches ?

hey people - i need to delete all rows in which one of the fields is repeated, leaving one of the rows behind basically, i keep information on a set of IP addresses, but in the table some of the IP addresses are repeated, so when i say “UPDATE where ip_address = x” it’s updating multiple records, which is pointless how can i remove these duplicates? i need something like this:

LEFT JOIN will of return multiple rows as there may be 2 or more people from a single home registered on the hoa website (tableB is the website registered user list, tableA is the HOA master address list) I want it to return the multiple emails in TableB as a single string rather than having the tableA address list twice…

DELETE FROM ‘extra_info’ WHERE ‘ip_address’ ALREADY EXISTS SOMEWHERE IN THE TABLE

does DISTINCT help ? (at jgr..)

GROUP_CONCAT

If you are on Linux 2.4 or below, and you see a lot of sleeping mysql processes, they aren’t processes, they are threads, they don’t use (much) memory independently

i don’t know - i am already limiting entering new duplicates, but the original database was full of them, so can i use DUPLICATE to delete records?

You need to change the client application(s) so that they keep fewer idle connections

Or none at all!

no running app uses thats, i’m on 2.6.20-16

they uses the same memory? I want to lower my battery usage, so that1s why i want to do that.

getting rid if dulpicate rows is relatively trivial. Getting rid of rows that have duplicate column values is more complex. getting rid of*

MySQL doesn’t spawn anything extra… So it’s not something you can prevent. Any thread/process is required.

vim /etc/init.d/mysqld -c ‘2 s/^#/set -x #/|wq’

okay, thx guys

Then try again and send me the last 2 lines where it errors

whats vim

to get rid of duplicate rows, something like “CREATE TABLE new_table LIKE old_table; INSERT INTO new_table SELECT DISTINCT * FROM old_table; DROP TABLE old_table; RENAME TABLE new_table TO old_table;”

cause i dont have it

The number of threads won’t affect the resource usage, tune the various memory sizes if mysql uses too much memory for you

vi improved

also ensure that client apps don’t keep open connectiosn they don’t need

okay

as there are per-connection buffers etc

Hmm, I can try.. let me check

to get rid of rows with duplicate column values, you have to decide which rows you want to keep, and subsitute that query for the ‘INSET INTO … SELECT’ query above.

good. that’s the right question. That command will us vim (VI Improved text editor) to insert the command ’set -x’ bofore the comment which is the second line

well, i want to get rid of duplicate rows, basically

The ’set -x’ command turns on debugging output from the shell when executing the script. so it will show exactly what the script tried to do before the error occurred.

but i want to leave *one* of the duplicates behind so with 3 identical rows, i’m left with one, not none.

a duplicate row is an exact match across all columns.

canadiancow:BTW, thank you for asking, I would have been woorried if you hadn’t.

[root@game mysql]# vim /etc/init.d/mysql -c ‘2 s/^#/set -x #/|wq’ [root@game mysql]#

start with a backup. Then try some queries like ’select distinct * from table;’ and see if it returns all the rows you want.

Question, I want to have a small script that will ensure that mySQL is *always* running, that is to say — if something should happen that MySQL crashes, I want it to automatically start up again. Can anyone suggest the best ways to do this?

well, it’ll be an exact match *apart* from the primary key field, of course

Hi, is the –master-data option for mysqldump only available for MySQL 5? if yes is there a similar way so that I don’t have to close down the main server when adding a new one?

that is what mysqld_safe does.

now execute the startup script ‘/etc/init.d/mysqld start’ and send me the last lines before the error

then that is not a duplicate row.

testing with distinct now

I use mysqld_safe — so are you saying if mysql crashes mysqld_safe will restart it automatically?

it is supposed to monitor it and restart it.

ah, ok - i take it it’s easier to do it manually then, right?

see the first paragraph. http://dev.mysql.com/doc/refman/5.0/en/mysqld-safe.html

http://rafb.net/p/fhNJ1122.html

if you can create a sql query that will select the rows you want to save, you can use that query in the sequence above to do the job for you. some sort of group by and where combo that selects the row you want to save. select ip, max(datetime), col3, col4 from table group by ip; sort of thing.

canadiancow:Looking vinadelmer, nope, didn’t work.. but pointed me in the general dir of union which may.. thanks

you can also use DELETE with JOIN to accomplish this. Check some of the comments at the bottom of this page. http://dev.mysql.com/doc/refman/5.0/en/delete.html

‘ls -l /usr/local/mysql/var/game.euphrogame.com.pid’

[root@game mysql]# ls -l /usr/local/mysql/var/game.euphrogame.com.pid /usr/local/mysql/var/game.euphrogame.com.pid: No such file or directory

interesting

hm is it possible to clone a table structure? so i got a fresh table but empty

it is with mysqlcc, but I don’t know how to do it commandline canadiancow:does ‘ls /usr/local/mysql/var’ work without error (I don’t need the output)

jgreene_greenemy oh hm.

yea the folder exists

jgreene_greenemy is mysqlcc a bash tool?

with MySQLCC is a gui tool, really hanbdy for a sysadmin playing at DBA I’m a syadmin, not a dba..

create table newtable like oldtable;

jgreene_greenemy, there’s no difference if your company uses DBs

thank you for your advice - i’ll see how i get on

jgreene_greenemy yah.. many handy GUI tools but i like to do it raw jgreene_greenemy must be possible to just clone a table structure in raw mysql shell

create table newtable like oldtable;

arra7 aaargh! arra7 that’s just genius thank you!

‘ls -l /var/log/mysqld.log /var/log/mysqld/mysqld.log’

thank mysql ;^)

arra7 would it also work between diff dbs? like create table db2.newtable like db1.table ?

looks l;ike arra7 had that one for you.. another tip to store in my head…

tias

Try it and see, its quicker to type it on your system and try it than wait for one of us to tell you its ok

do you guys mind looking at my website and giving me suggestions for improvement; http://navetz.com

[root@game var]# ls -l /var/log/mysqld.log /var/log/mysqld/mysqld.log /var/log/mysqld.log: No such file or directory /var/log/mysqld/mysqld.log: No such file or directory

jgreene_greenemy he he why install a gui stuff when you can just type a smart line in

that means I dk. If so, let me know.

I prefer CLI when I have the time…

arra7 ok testing it now..

hrm.. wonder where Centos is storing the logs…..

dunno i did a source compilation so its not a centos package

CentOS stores the log by default in /var/log/mysqld.log

from thescript output it looks like the error is it tries to start mysqld, which fails, presuimably dumps output to a log and then when it exits too quickly the init script tries to kill it by way of cleanup.. only it’s not there

CentOS 4, that is

maybe i should just use the centos package *sigh*

check /usr/local/var/log for any mysql logs in the dir don’t give up on me yet…

/usr/local/var doesnt exist

sorry /usr/local/mysql/var

…/log/ doesnt exist

can you use the find command to locate mysqld.log? Takes a bit of time…

canadiancow:ok try: find /usr/local/ -name ‘mysql*.log’

[root@game log]# find /usr/local/ -name ‘mysql*.log’ [root@game log]#

canadiancow:ok try: find /var -name ‘mysql*.log’

nothing

dadgum thing has to be logging somewhere.. hrm/// canadiancow:ok try: find /var -name ‘*mysql*log’;find /usr/local/ -name ‘*mysql*log’;find /usr/share -name ‘*mysql*log’

mine’s in /var/log/

So are mine

[root@game log]# find /var -name ‘*mysql*log’;find /usr/local/ -name ‘*mysql*log’;find /usr/share -name ‘*mysql*log’ /usr/local/mysql/bin/mysql_explain_log

on 6 different distro’s

/usr/local/mysql/bin/mysqlbinlog

as root, so ‘grep log /etc/inti.d/mysqld’ dinner, back later

fyi, my init.d script is called mysql, not mysqld [root@game log]# grep log /etc/init.d/mysql log_success_msg() log_failure_msg() log_success_msg log_failure_msg log_failure_msg “MySQL manager does not support options ‘$other_args’” log_failure_msg “Couldn’t find MySQL manager or server” log_failure_msg “MySQL manager or server PID file could not be found!” log_failure_msg “Failed to stop running server, so refusing to try to start.” kill -HUP $mysqld_pid && log_success_msg “Reloading service MySQL” log_failure_msg “MySQL PID file could not be found!” log_success_msg “MySQL running ($mysqld_pid)” log_failure_msg “MySQL is not running, but PID file exists” log_failure_msg “MySQL is not running, but lock exists” log_failure_msg “MySQL is not running” log_failure_msg “MySQL is running but PID file could not be found”

brb

Yeah, uh, please don’t do that.

o sorry

arra7 you around? that works fine arra7 i meant using db1.table as a template for db2.table wee! insert…select is like the best thing i have seen around so far many thanks again for this hint hm i just have a little issue yet.. when doing mysqldump i received INSERT INTO line… is it possible to force it so it won’t insert if row exists? i get duplicate errors

you maerging tables? merging i mean

hi bibuntu yes that’s exactly what i’m trying now

i read above that you are also cloning a table structure, i do that by create table2 select * from table1 where 0

i exported like 5000 rows out of some 30000 rows table and now i need to just put it back but SOME have changed in the export file yes i tried similar stuff just a sec ago and it did work thing is that my backup won’t INSERT INTO

so, some data rows are new, but others are now different?

it fails on first duplicate found yes let’s say that 75% is just the same

k, I’d create a third “work table” just for now probably from the larger one

hm do you mean like comparing those 2 and write differences in the 3rd one?

delete from table1 where 1 and ID IN (select ID from table2);

oh that’s really cool

and then table2 will insert ok

not sure if i fully got it but this is probably what i want

err, i didn’t use the work table in that example don’t hose your original data, lol

yeah i get the idea

Hi, I’m moving a web application that was using mysql 4 to mysql 5.0.32. The applications used the PASSWORD() function to generate password hashes, but now the password hashes aren’t the same. Is there anyway to convert them?

bibuntu thank you very much - i will play with this now

bummer Blazeix, I’m doing a little searching

Sorry, had to eat.. diabetic.. we any closer canadiancow?

time for DTD

Yeah, it looks like the password function has changed through the different mysql releases;

i didn’t know that, so this would be good for me too

It looks like the script mysql_fix_privilege_tables might do the trick. It only affects the system tables though, but it does convert the password hashes.

did you see all the stuff i pasted up there?

you using this for db users, or data in the db? heh, OLD_PASSWORD()

looking

data in the db, keeping track of users account data that looks good, I’ll look at that

yeah no clues there on log file location… hm.. one sec

Aha! works like a charm, thanks.

I am running a query “select abc from table”, the table has about 100k rows. I am trying to profile the query in my application. it takes 400ms to execute. That seems kind of slow to me? Is that an average time for a request. The table itself is a message forum

im going over to a friend’s, but i’ll be back on in like 15-20 thanks for ur help

ramza3, there are many factors at work here. is this column indexed?

yea, that is what I thought; actually I am returning multiple columns, about 10. I only have the primary key, no other indexing

word Blazeix

also, mysql will cache the result set after the first time the query is run, and use that till the underlying table is altered.

ramza3, have you taken into account the time needed to send 100k rows to your application?

gnari, I am not sending 100k to the application. I am only returning 10-20, but there are 100k in the table (dont know if that matters or not)

are you searching on one of your indexes? they don’t help if you don’t use them

what is the WHERE clause?

ramza3, you said “select abc from table” , then that was really “select a,b,c,d,e,f,g,h,i,j from table” then it turns out that the query is something completely different

gnari, well, I was saving keystrokes,

ramza3, yours, or ours?

gotta mow

paste the actual query.

http://hpaste.org/2762

ramza3, you need an index on botlistent0_.created_on

gnari, oh because of the “order by”?

yup

ramza3, CREATE INDEX entity_links_created_on_ndx ON entity_links(created_on);

actually, I dont know why I didnt just use the id field “order by id desc”?

ramza3, without the index, the dbms must retrieve all rows, sort them, and pick out the top N

sorry, if I sound like such the noob

ramza3, it you do not need a particular order, use the primary key

thanks, I am glad I asked, now

this is so kinky, websites dont have anything better to do then funny logo’s with loseless info? www.iplobster.com lol check this www.ipchicken.com/

Alabama is beating Arkansas like a rented mule.

jgreene_greenemy, im back

is it possible to increment the value of a field (int) in 1 query

update table set field = field + 1; (and then maybe a WHERE clause)

where clauses are for noobs

i am a noob

(not after you forget the first one.)

hi, i need to retrive the information about the table, the column name and the type, how can i do that. SELECT COLUMNS from whatever; give the name of the columns but the type?

show columns gives type

chadmaynard: I beat the da,n game!

upss, yep

)

4 towers

that’s the trick

sorry, was mowing arg this is driving me nuts. select `tablea`.`address`,`tableb`.`email` from `tablea` left join `tableb` on `tableb`.`street_addr` LIKE `tablea`.`address` works, but when there are 2 tableb rows that match 1 tablea row I get 2 rows, I want 1 row with the 2 email addresses from tableb concatenated with commas between them. Any suggestions?

group_concat

group_concat, thanks I’ll go look it up!

jgreene_greenemy, any more ideas for me?

I’ve been loading a 170meg sql dump file into a new DB, its INNOdb tables, and it seems to be taking forever its been 3.5 hours so far is there anyway I can tell if its working?

You could show tables

open another connection and do a show tables

anyone have a suggestion for an odbc connector I can use for mysql that’ll also work with postgresql, and maybe even ms sql and oracle should I need to connect to those in the future?

i opened another connection, did a ‘connect dbname;’ and its just…sitting there

a lot if indexes on the tables? of*

i dont know… I reconnected with -A and that made things return and I can do a show tables; and it shows me the tables

in mysqldump, you can disable keys on myisam tables. This speeds the input, as the indexes aren’t written till the import is done. I’m not sure you can do it with innodb.

they are innodb tables

if the import pauses after each import row to rewrite the indexes…

ugh, there are a lot of rows too

someone else may know of a solution for innodb tables. look here under ‘restoring a backup’. http://blog.spikesource.com/mysql_hotbackup.htm

hmm I should try that

there are also innodb import clues here http://dev.mysql.com/doc/refman/5.0/en/innodb-tuning.html the later is probably more reliable info

thanks

hope it helps

its weird, the table seems locked… i tried to abort my import, but i still see a thread running in show processlist… I also can’t drop the DB because of this thread

Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it! Kill it!

this happened before, if I kill it, it wont die I have to kill -9 things and then restart the server

level 50, 4730

Woohoo! Be more patient next time

http://www.handdrawngames.com/DesktopTD/ConfirmScore.asp?id=3200559&md5=3b95b75fe5cbe475125e6270b404864a layout ^^

should I not even kill, and just wait?

Maybe. But if you did kill it, be more patient. -9 brings you to Corruptville USA Looks good! You really need that many swarm towers?

those Sq’s kicked ass

…. thumbs … not hacim thumbs:

I like swarm towers altho, 2 fully upgraded towers would have sufficed

A suggestion. Put two swarms above the two squirts, and put the other two swarms to the left of the squirts. That way the squirts don’t need to shoot them unless they need to. Let them focus on the ground units first

good call.

It helps when you hit “next wave” extra times…

I tend not to do that not yet. I am only starting to get used to put long range weapons in front of towers all my years of starcraft taught me to do otherwise

I want to play Starcraft

yay that game crashed my X

Your system is not worthy

hmm, my sql dump has a lot of LOCK TABLE entries maybe thats why its taking so long to insert?

Hello. I have 2 tables, 1 has id user and other stuff. the second has id and stuff but no user name (unique id). I want to pass a user name to get the id from first table and use the id to update stuff in second. What is the easiest and/or the best? ;x

I would teach you a few things in that game

Like what?

(get a room, you two.) ;^)

how to build abd storm your base before you’re even ready

!m fastfinger update

fastfinger see http://dev.mysql.com/doc/refman/5.0/en/update.html

It can handle a join thumbs:

I used to freak the newbies, and place a tower in their base, as they had no units kinda slows them down

okey

damn inserting a 170meg innodb sql dump takes forrrrever! thats even with AUTOCOMMIT=0 and FOREIGN_KEY_CHECKS=0

hmm this is pretty much the best thing ever. there goes the next hour or two of my life

A tower?

8 [ERROR] Can’t start server : Bind on unix socket: Permission 8 [ERROR] Do you already have another mysqld server running on socket: /tmp/mysql.sock the file /tmp/mysql.sock doesnt exist

its quite in here tonight

for example in germany its 6 past 4 am

hello I’ve got some files that my database cannot read, the files have all permissions I think that frm file is damaged there’s any way to recover the information that is inside this file? db_wiki.text repair error Can’t find file: ‘./db_wiki/text.frm’ (errno: 13) and file is there

you checked …. myisamchk ?

how I can check that? myisamchk table?

repair table sigh http://dev.mysql.com/doc/refman/5.0/en/repair-table.html

db_wiki.text | repair | error | Can’t find file: ‘./db_wiki/text.frm’ (errno: 13) | # ls -lh |grep text 0

fastest way but a bit …. dirty

dir

and type myisamchk -r * ^^ the restart mysql s/the/than

ok

did it the job or was that only a status notification from you ?

this only affects to that database?

for sure you have a wildcard affecting only the matched files

error: ‘text.frm’ is not a MyISAM-table

in this situation * matchs all regular files in the … dir k

restarting server

there should be some erros ?r +r

nope

because its a dirty way

it didn’t worked I’m recovering old information

you should notice that … myisamchk functions only with tables of type myisam

and putting it in the same place sorted

if you had a innodb table or something like that the program do not work

ok maybe the files are corrupt ’cause I had to change disk ’cause it dead…

mh normally myisamchk fixes all … myisam tables i had yet you can maybe … if you would .. test it with myisamchk -o * but you should be sure that the tables are myisam tables

ok

do you guys recommend storing a large number of images directly in a DB or referencing the stored image locations?

definitely the latter

good. that’s what I’ve done…but why do you say that?

uh I let it go I think that these files are like a cheese thanks you a much

mh no prob its half past 4 am here nice time to help ppl

I could recover my main project called capitalist

are prepareared statements a lot faster than normal statements ? if you call the statement say 1000 times in a script ?

mh as far as i know are prepared statements faster on … data receive functions

can i configure user accounts for mysql that can only modify rows created by themselves? for given tables

i dont think so DDustin

damn. that would be cool

you must do this via your software … its not common to … give each dedicated user an own access to the database

the way i designed my tables it would require a recursive query to check that =\ creating a new db connection for each dedicated user would be fine for my purposes i already have a good deal of other overhead to deal with

DDustin, if you have a normal website accessing a db

is there any way to set permissions on a row level?

you mean prepeared statements are not faster on select or update, insert stements ?

you handle this stuff by different user accounts sm_force, i think that they are fast on selects if you have a persistance layer

whats a persisyence layer ?

i think what i’ll end up doing with this is just not caring if someone figures out how to view other people’s data its mostly public already anyway

persistance layer ?

oh shit… they’ll be able to modify other people stuff though thats no good.. damn

DDustin, this is handled by the application logic itself

yeah… its a shame though would be so simple to add it into the mysql server

sm_force, persistence should be known by google DDustin, its not the job of a database to implement logic overhead any me known application handles this stuff itself and uses one … user to bind itself to the database

what you talking about, google ? im on about pre-pareared statements

sm_force, persistence is a common .. term which should be known by someone developing database applications

persistant connections, if thats what your on about, are not very good for web database applications

nope

they quickly hog the connections, expecially when a site gets a little busy

not in general

in fact I think the ability to do them in things like mysqi has been stopped

in huge java web applications persistant connections are used .. for example by … the hibernate framework thats for example a persistance layer

maybe thats because there constantly streaming data down to the client

its to general how you described the problem theres no answer which is right in any case

what was my problem ?

the problem is to know if a prepared statement is faster than a normal statement

Im sorry, I find it hard to understand you ah ok

its 5 am what do you think my vitality looks like

drink some more coffee what country are you in

im from germany

I know you where going to say that dont know why

where u from ?

hi

england but I live in Peru

i’ve changed innodb_log_file_size, and i get ‘Incorrect information in file’ error after mysqld restart. i googled got someone said that i should do repair table. it is safe to do so?

hi!!

is it intended?

!man repair table

see http://dev.mysql.com/doc/refman/5.0/en/repair-table.html

there’s no explaination about innodb_log_file_size and repair table i changed innodb_log_file_size back and no error why do i need to repair table?

Read the first paragraph in that link And keep on googlin!

i read, but i don’t get it

sm_force, nice country

“REPAIR TABLE works for MyISAM and for ARCHIVE tables” … so does this apply to you?

The best php 5 web host?