@# Quotes DB     useful, funny, interesting





Google
 
Web www.quotesdb.info
Undernet  |  EFnet  |  Quakenet  |  Freenode  |  Dalnet  |  Ircnet  |  Galaxynet
Page: 1 2 3 4 5



Comments:

<0> it's part of the total cost of ownership for a wannabe free database
<1> and no, the free version doesnt have many features it used to lack
<2> there's a non-free version?
<3> i thought the main features it once lacked were: stored procedures, views and triggers?
<3> doesn't it now have all three?
<1> and transactions
<2> and subquerie
<0> and....
<2> and constraints
<2> and relationships
<1> and reliability
<3> wait, the transactions and subqueries are definately there now
<3> i'm not defending mysql, just trying to get an accurate picture of where it's at now
<1> well itll get you banned shortly
<2> i remeber someone in #hardware
<2> "mysql is so cool, it has a tool to recover data"



<2> basically, mysql sometimes ****s up the data
<2> and they have a tool to fix it :)
<1> giddy, should a database let you insert 31-FEB-2006 as a valid date?
<2> does it really do that though?
<2> haha
<1> yeah thats why that stupid topic was up for a month
<2> i thought it was a joke
<0> what happens when it's in the db? does that get converted to like March 3rd?
<1> depends...
<1> sometimes it does sometimes it doesnt
<1> all depends what function you use to pull it out or modify it with
<0> heh nice
<3> Jarett: why not leave it up to the code to validate something like that?
<2> haha
<2> that's why you are using mysql not some real DB
<3> well i don't claim to be an expert, nor do i claim it's the only way to do it -- can you explain?
<1> the people who make mysql are stupid
<1> and dont know the basics of calendar logic
<1> so they check for any day under or equal to 31... as a valid day of the month
<2> not only that. you have to be able to trust your data
<1> that the most complex part of their validation on dates
<1> they also think 1 = 2 in some cases...
<2> and their date searchs probably ****
<1> and null = null in most cases
<3> hmm well i've never tried it - you can insert a 2006-02-31?
<2> because they seem to be handling dates as strings
<3> ok
<2> try it
<2> you think we have mysql installed?
<3> trying it right now, heh
<3> naw that's not true, at least for this case: update books set last_updated = '2006-02-31 11:11:11' where id = 1563
<3> result is: 2006-03-03 11:11:11 for some reason
<3> but i guess your point is.. it doesn't reject it?
<3> whereas something like oracle will?
<2> ofcourse
<2> because that's an invalid date
<3> well hey i'm not a big oracle user, so it's not that obvious to me
<1> i dont know any other database that would allow it
<1> plus if you insert a day like 999-03-99 11:11:11
<1> itll go null
<2> where?
<1> it wont reject it
<2> eeewz
<1> but it will reject 999-03-01 11:11:11
<1> which is a perfectly valid date
<1> i guess if its before the dark ages... mysql doesnt want to know anything about it... cause thats where mysql live
<2> you seem to know alot about mysql
<2> are you an under-the-table sql user?
<1> i use or have used almost ever rdbms and many oodbms
<3> update books set last_updated = '999-02-31 11:11:11' where id = 1563; sets it all to 0000-00-00 00:00:00
<1> hating something is a bigger reason to know its weaknesses
<1> giddy, yes thats their "null" date
<3> ah
<2> muhaha
<2> null date
<3> wait, is there such a thing in oracle? a null date?
<1> since null was too hard for them in dates :P
<2> haha
<1> a null is a null no matter what the datatype
<2> null means "missing"
<3> well, it shows up as "null" in oracle?



<1> no
<3> null means "nothing" and not zero
<1> no
<1> its just null
<2> stfu
<1> nothing is something more than null
<2> null means not present or missing
<1> no you can insert null
<1> which would be there
<4> http://www.tmcm.com/comics/webcomics/020_shakespeare
<1> null exists.
<1> but isnt
<2> means you would set it to missing
<4> oops sry wrong window
<3> hmm null is there but not there
<2> lets say you dont have someone's phone #
<2> you don't have it, therefore it is missing, therefore it is null
<2> null is not a value
<2> it's a concept
<1> depends on the dbms
<1> i only care about oracle's null
<4> there's NULL, a zero length string, and whitespace ...does everyone know the difference YET?
<1> this index build has been taking sooo long
<2> jare: cause butter fingers
<4> well you just got rid of all your stats
<1> no this is a new index
<4> oh
<1> ill have to stats it later thatll take 2 hours
<2> jarett: will you teach me whoracle if i ever need to?
<1> with 2%
<1> no
<2> why
<1> oracle must be learned it cant be taught
<2> huh huh
<1> and i charge 7500 a week when i teach it
<2> yes but i want to bug you about every little thing
<2> when i am the process of learning
<2> by the way
<2> is there an insert/update statemenet in oracle where it will update instead of inserting if a record exists
<1> yes
<1> a few of them
<2> which ones
<1> merge is the newb option
<1> aka upsert
<2> upsert eh?
<2> how does that work
<1> very well
<2> does it check distinct rows
<2> or
<1> try it
<2> does it check if you p***ing a PK or not?
<1> but it depends how much data. im about to start recreating tables :P easier to create table as select (work here) and rebuild the indexes and then blow away the old than insert/update on 50m of 500m rows
<2> oh yah
<3> i was reading that I ought to move away from mysql if i want to track change to a table (audit) -- somehow it's built into oracle/postgres
<3> is that a pain to get going?
<3> but sounds good that it's built in
<1> can you afford one of the pay versions of oracle?
<1> not sure how easy auditing is to enable in the free oracle
<3> i was thinking of instead, just making a "log" table of each table i want to audit changes
<5> that would work.
<1> well its your code
<3> so before each update/delete it will log changes to the log table... then i can use that table to generate reports on what changed... but problem is it's on a per-table basis
<1> you set the rules
<5> a series of meta records.
<1> what about selects?
<1> if you need to be PCI, HIPAA or SOX compliant you need to log selects too
<3> well for this, i think i'm just interested in changes to the data
<1> but for small apps like this its not that important
<5> Jarrett: that must create a terrible amount of logs.
<3> for PCI, HIPAA or SOX they require you log the person/user too right? is that essentaillly what they are about? basically log everything to hold ever user accountable?
<1> it gets worse when you need to log everything going over your network :P
<3> well, log and to be able to retrieve data in a reasonable amount of time i guess
<1> i sort of wish my company had to be held to those standards


Name:

Comments:

Please enter the result of the sum 63 + 46 (to avoid spam):






Return to #sql
or
Go to some related logs:

glycerine suppository hot -health -doctor
#politics undernet jews
#bsd
iosoccer source
robtv, candax energy
rope tied around my penis
what do you mean domain controller *
#flash
#nhl
#politics



Home  |  disclaimer  |  contact  |  submit quotes