| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Comments:
<0> and i've heard .net 2.0 is the only way to fly
<0> templates and incompletes and all kinds of joy
<1> ChillAxen: um.. not sure
<2> get what i mean though
<1> unless you meant bitshift by bitpush
<2> yep
<1> but yes, you'd bitshift the larger masks into position and or them with the rest to get a composite bitmask
<2> so as i build the bitmask it will first keep adding to say $output[0] till its full then start adding to $output[1] and so on
<2> and on SQL i can jsut do a foreach of the bitmasks doign a push of bits and oring
<1> so like 1101 and 1001 as two masks, to put them together to get 11011001 you'd 1001 | (1101<<4)
<1> ChillAxen: i'm going to go ahead and say sure
<2> mySQL mas is 8 bytes, which is 64 bits right?
<2> max'
<3> what do you need to use 64bits for?
<4> yea
<2> lol dam i explained that like 100 times fed. I have a list of on/off flags so im converting all these flags to a simple integer
<2> then from there it would be WAY faster for mySQL to just compare 2 numbers
<2> then say have each onoff as a field to check
<5> building a relational database using bitmasks is dumb
<3> don't call me fed >:|
<4> you can't sub-group all these on/off switches into groups less than 32?
<1> ChillAxen: yes that's 64 bits
<6> is there a way to have a tiff image on my hard drive to show as a valid resource with imagecopyresampled
<2> kicken: i have no clue how many they will add later in life to this app
<1> kicken: i've tried suggesting that several times now
<2> which in essense isnt really that good either
<6> when the img already exist
<2> if they ever go over 64
<2> then it will outlive it's logic ;/
<4> i was just about to say that
<2> but can you think of a better way?
<1> ChillAxen: depends on your means of grouping them
<1> if it goes over 64 bits, then your outliving your logic anyway by storing them all in one mask in the db
<2> the example of this specific part is a restaurant database
<1> the points is that the groups themselves should have their own db field as well
<4> ChillAxen use your dreded joins? with the proper indexes and such it shouldn't be a problem.
<2> and this is just storing different types of foods they serve
<5> TheBomani: hmm looks like gd doesn't like TIFF
<1> doing it with a db requires simply a linking table, and a 2 joins tops
<1> but it's a lot slower
<2> kicken: idk... if theres millions of hits a day
<2> wouldnt that be ALOT slower
<1> well, not a lot
<1> oh i dont know
<1> i'd have to bench mark it
<2> two: yeh i know how to do joins with it, but im looking long term
<5> ChillAxen: using a bitmask in a relational database is going to bite you in the *** later, use the model the way it's supposed to be used and you'll be happy
<6> pizza: I know. Thats my prob right now
<2> if it had millions of hits a day, this comparing 1 number instead of a **** load of left joins
<2> it would be smart
<1> it's not a ****load of left joins
<1> it's 2
<1> always 2
<2> two: it then has to compare on every item in table 1
<1> you mean for each record?
<0> i don't see anyone suggesting 33 columns of booleans yet
<2> such as SELECT t2.* FROM restuarants as T1 where T1.id = T2.restaurant_id
<0> why isn't anyone suggesting that yet
<1> because you should not be enumerating multiple items as multiple fields in a table
<2> jed: cause this is just a array they can customize
<0> oh, customizable
<1> jedzilla: because that's worse for scalability than ANYTHING else
<2> yeh i did that already for bar opening times based on m-s
<2> and i was hesatant to do it
<2> brb
<0> so between lack of scalability and complete udder disgusting code
<0> the consensus is, "scalable!"?
<1> uhh, 32 boolean fields (which would likely be stored as ints in the db anyway, wasting a ****ton of bits) would be even uglier to code
<0> mysql has a boolean doesn't it
<0> that's actually a bool
<1> where bool1 = a OR bool2 = b OR bool3 = c, etc
<2> not really jed
<2> just a int(1)
<1> i doubt it. prolly 4 bits
<0> oh, poo on it, then
<2> which still isnt a true Bool
<1> er, 4 bytes
<0> no wonder i don't use it
<2> cause a bool is just 1 bit
<2> well two, the do have int(1) which is closer than that
<2> but still not a true bool
<0> $str = "\Unow this is just ugly\E";
<0> why am i learning perl? anyone?
<2> lol
<2> idk
<1> we hardly ever use \U and \E in perl
<3> if a bool is just 1 bit, then when it's false, does it just have no value?
<0> it still exists
<2> why when i went ot college did i have to learn cobalt -=/
<1> so?
<0> kinda like, the klingons are in the federation, but they're still ugly
<1> there are a ****ton of functions in php that exist that you'll never use
<0> i love that '****ton' word
<0> is that like 'bogon?'
<2> fedt: idk... isnt 1 bit either 0 or 1
<0> 'bogosity' and 'quantum bogodynamics'?
<4> I use tinyint(1) for bool's in mysql
<1> heh
<2> kick me too...
<2> but still not a true bool
<0> quantum ****tondynamics?
<1> ****ton++
<3> yeah, that's right
<2> cause it can still store 0-9
<2> so a bool of OFF is 0
<7> me four
<2> and ON is 1
<7> just use it
<2> or NO = 0
<2> YES = 1
<1> tinyint is 2 bytes, right?
<4> could use an enum :)
<3> or SOMETIMES or ALWAYS or NEVER
<3> or MAYBE = .5
<2> Fedt thats not a bool then
<3> :) i know
<2> =]
<1> he's being an ***, ignore him
<2> its cool
<3> i would just love to say php generate a "maybe..."
<3> *see
<1> no, it's pathetic
<1> we're gonna take him out back and beat him with a hose
<2> idk i might just store fricken 4 different INT fields
<2> an d if they go over 32*4 **** them
<1> i wonder how hard it would be, sql wise, to store them in a binary string
<2> i dont think mySQL has that
<1> might require a quick UDF to handle the data manip
<2> cause thats waht i wanted to do lol
<7> ChillAxen- you are storing "options" for each restaurant?
<1> mysql has varchar, no?
<2> yes varchar
<2> they even have a varbinary
<2> but thats different then binary's
<1> that's all you'd need
<1> it's data that's stored in binary form
<2> Two-bits my idea was to take the same method im doing now
<2> and store as a string
<2> as apposed to int
<1> using the quick string functions to get each char would effectively get each byte
<1> not as a string of 1's and 0's though
<2> care giving a quick example of what you mean?
<1> build up a byte in the bitmask, chr it, add to string
<0> Two-Bits, do perlmongers use qw() much
<2> olinux: im really concentrating on this because it will be used for LOTS of different things
<0> it seems like it could be a timesaver
<0> @zip = qw(foo, bar, bing, bang);
<1> jedzilla: for exports/inpurts mostly
<0> er, foo bar bing bang
<2> such as this restuarant tab has a list of optional food types they serve as one and another restuarant types
<1> use Foo::Package qw(funcs that you want to use);
Return to
#php
or
Go to some related
logs:
choooookie
Seinfeld season 10
#linux
#india
#india
islamabad anties
Angad ramayan
#india
#php
#linux
|
|