Mitani: I’m running a select statement on a few thousand rows. i get lot’s of NULL returned for a column, status.
Tesauro: I’d like to update all these NULLs to ‘Unknown’, but when i run the query it doesn’t find them. I’ve tried the following
Grensky: UPDATE table SET status = ‘Unknown’ WHERE status = NULL;
Kornn: After one of our raid disks crashed but didn’t get rejected I’m having some serious issues with mysql constanlty crashing, I know what to do with MyISAM tables but no idea how to recover/check InnoDB stuff to get this to stop. Logs: http://pastebin.com/q8nUSqZG
Haifley: Aias: WHERE status IS NULL
Balak: Hi, is there a way to find out what my binlogs are doing, for some reason my binlogs have increase in size more then normal but i cant work it out why
Heinlen: Lanox: mysqlbinlog + pt-query-digest
Cauazos: Bert_2: you mean a raid5 disk got damaged/broken but your raid controller/software didn’t realise this was the case?
Mannon: Sounds like there’s a pretty high likelyhood that your datafiles were damaged. I suggest you wipe the whole thing and restore from backup.
Kuerbitz: If that’s not an option, gradually increase innodb recovery level until it recovers whatever there is to be recovered. Do read the appropriate do***entation thoroughly, though.
Auteri: Ok, bit of a noob question, but here we go: I’m trying to find the first blank field in a column and write data to that field. Would I do this using INSERT and SELECT?
Maasch: Presumably NULL or empty string?
Rebeles: Right now I’m using LENGTHsku 2
Immen: Vegivamp: i was using it today but i couldnt not work out what was going on
Semprini: I will try pt-query-digest and see how i go. I did notice that one database had lots of retrys
Marlett: But i am not sure if that is a problem or now
Mishra: Yeah, plain mysqlbinlog is a bit unwieldy :- pt-query-digest will default group by statement fingerprint, so you can see which statements are being overused or being horribly inefficient. You can also tell it to group by database, user, and whatnot.
Wriston: Sonny_Jim: why 2? I’d think you want where stu = ” or stu IS NULL
Wineland: Vegivamp: no, software raid1 mdadm
Rutan: Vegivamp: with force recovery I could dump stuff, so I’m now indeed going to drop everything
Rabenhorst: Same problem for raid1, really – if your software doesn’t realise your disks are no longer in sync there’s no telling which version of a given block you’ll get when reading
Isa: Although writing should probably go both sides and “fix” that bit of corruption, under the simplistic ***umption that you didn’t base the updated block off the wrong version of the old block
Mullings: Vegivamp: In some of the fields it has ‘r’, hence the 2
Barbadillo: Really I should go through and strip them
Volmar: Ideally you’d grab the backup from right before the corruption happened, ***uming that is a defined point in time and not horribly long ago, Bert_2
Sollberger: Sonny_Jim: that would be a better thing, I suppose :-
Mumm: First harmonize the data, then compare it to a summer’s day
Pittard: I have a select query from a single table. Select ID, VALUE from table where ID=1 or ID=2, I want to split the values into new columns depending on whether they have an ID of 1 or ID of 2, any thoughts?
Copa: Because at the moment I have a row for each MASTER ID, where ID=1 or ID=2
Weinreich: I want 1 row for each master id
Pelham: It’s really hard to explain, i think i need to pivot the table but not sure :/
Keiter: Giza: you want new rows, or new columns?
Zurovec: New columns which will reduce the number of rows by half
Keiter: Giza: why do you want separate columns?
Keiter: Giza: you know what, just show the results you’re expecting.
Bieberle: At the moment i have master id unique id name id =1 or 2 value