| |
| |
| |
|
Page: 1 2 3 4 5
Comments:
<0> :p
<0> <3
<1> Arrakis : Do you mean "you continue" ? :-P
<2> i would suggest, however, that re-seeding the random number generator in ms-sql with current time to the millisecond each time it's used will be a good start
<1> Yuppie : yeah, it was a suggestion made on one web-site but he reported that it didn't provide sufficiently random results for some reason - perhaps because calculations are just that fast, I'm not sure
<0> are you going to take some ****tards word for it ?
<0> lol
<2> so just use an external rnd generator - write one in c++ and call it from ms-sql
<2> or, as i suggested earlier, use one done in c#
<2> givne that ms-sql2k5 supports that now...
<1> if I must I'll do it post-query like you're suggesting... it's just that it takes more processing power to go through retrieving a larger set of all possible incorrect answers, then running through them with a random number generator in post-query code.
<0> you're talking about creating x amount of GUIDs ?
<0> as opposed to a seeing a generator?
<2> not if you're correctly normalised - that way you're only returning a list of answer ids and picking randomly from that
<2> rather than pulling back the entire list of possible answers' text
<2> or, as i keep saying, write something in c# to use within ms-sql2k5
<1> yeah, I might have to end up doing that then unfortunately. I was hoping it wouldn't be a complicated and time consuming effort is all
<2> well, it won't be
<2> prolly take 5 mins
<3> Ok, I didn't realize you could code sprocs in .NET languages with mssql 2k5. That's a huge.
<2> Drk`Angel : yup
<2> Drk`Angel and anyone else interested : http://download.microsoft.com/download/f/a/a/faace3f4-03c1-4455-8acd-a8d7ad959ab3/SQLServerExpressSamplesOverview.htm
<2> second item called "sql server 2005 samples"
<4> you must've been some trouble to get to that page...
<3> Oh man, I may have to actually try to convice my boss to upgrade.
<3> Crap. :)
<4> do it do it do it! :)
<4> well i convinced my boss to let me swich from vb6 to c# for my next project ~
<3> Working on moving some data out of a freaking ancient Oralce 7.2 database to something newer so maitenance is easier until we can replace the whole system.
<4> Drk`Angel:: oracle to something newer? gosh, i have to move some old Foxpro database in DOS that was created in like 1970s
<4> to MS SQL 2000...
<4> but since i am leaving the company, i don't want to worry about it
<3> d2d: Well, I didn't mention the ancient dos based system using clipper and dbase that we're migrating to msde :)
<4> hmm.. to msde... that sounds... worse :)
<3> The place I'm working now basically had no IT department for about 10 years at one point, so all of their tech is REALLY old.
<1> Have they changed the syntax of a view in 2005? suddenly I'm getting an error
<3> I feel like I'm working at a musem sometimes.
<4> i see.. kind of in the same situation i am facing here
<1> create view viewName select * from blah right?
<5> as
<5> i think that was last changed in 1992
<2> indeed
<2> create view blah as select * from table
<1> actually I had the "as"... accidentally didn't type it
<1> still getting an error - weird
<2> maybe you should tell us what the error is?
<2> just a thought
<3> How would that be useful? ;)
<4> Eeker:: you can paste your query on the site listed on #sql topic
<1> Incorrect syntax near the keyword 'view';
<4> what's your "viewname"
<4> well just paste the darn query darn it ! :)
<1> http://sql-servers.com/nopaste/?show=334
<2> wtf?
<2> a stored proc that creates a view?
<4> wth
<1> maybe I misunderstand the purpose of a view
<2> i think you do
<4> Eeker:: if you really want to pull that off, you can use hmm... dynamic sql
<3> I'm thinking yes.
<4> but i don't think you shouldn't ...
<4> s/don't think/think
<1> I guess what I need to make is a temp table
<1> that's it
<3> What, exactly, are you trying to do?
<2> man, that is *so* broken
<3> Other than just stabbing in the dark?
<5> or uhh, just remove the whole create view crap and leave the selects and you're on the right track
<1> I'm trying to find a simpler randomizing solution than going and writing my own randomizer. Those two select statements individually properly give the two components I need
<1> but if I union them it doesn't work
<2> you don't need a temp table, just a union
<5> what doesnt work if you union them?
<3> Why in the name of all that's holy isn't your first query "select @elementid" ?
<2> select id from table where id = @id union select id from (select top 3 id from table where not (id=@id) order by newid())
<2> is how you do that
<2> Drk`Angel : good point
<5> because its a parameter
<3> So?
<2> phargle : doesn't make a difference
<5> whats the significance of having a select @elementid?
<3> I mean, maybe he's verifying that the row exists ... but based on the rest of the query, I really doubt it.
<1> my first query's not select @elementid because I want more than just the elementid
<2> the only time you can't select a variable into an result set is when you're using mysql
<1> it's a foreign key in another table
<3> eeker: All you're returning right now is the elementid anyways.
<2> Eeker : then why are you not selecting it?
<2> s/it/them
<5> makes no sense..
<2> Eeker : anyway, that's neither here nor there
<1> in what you guys see, yeah - sorry
<2> select id from table where id = @id union select id from (select top 3 id from table where not (id=@id) order by newid())
<1> it used to be a *
<2> is what you need
<1> I actually need the sub-select twice... on the first one too since there can be more than one right answer
<1> as I said, the id is a foreign key
<1> but otherwise, I'll give it a shot
<2> well, you never said before that there could be more than one right answer
<1> sorry
<2> otherwise i'd have given you different code
<5> can an answer be both wrong and right?
<3> You could just write a function that returns the result set too.
<3> No reason to toss it into a view or a temp table.
<2> indeed - my code was a replacement for the creation of the views - i should have been clear about that
<2> right - biab
<1> his code is giving me an error
<6> why aren't you randomizing your result set in your consumer?
<5> i recommend using randomizer logic that makes the correct answer always "C"
<5> it helps with testtaking and all
<1> MuteThis : my consumer?
<6> yes, the code that is requesting the result set
<6> it should randomize
<6> for example, why have two procs, spGetResponses and spGetResponsesRandom
<6> just have one and let the consumer randomize if necessary
<1> I'm sorry - I still don't understand what you mean by the consumer
<1> you mean in my C# code?
<1> in asp.net?
<6> yes
<1> that means I'd have to return the full set first...
<1> but my query limits to only 4 total records
<1> I don't wanna have to return all those records first and then randomize them
<6> sounds to me like you should cache the table and then pick a random four from it
<5> random 3 + 1 correct answer
<1> ok, I finally got it
<1> yep, exactly what I did
<1> MuteThis had such a simple answer and that was what I was going for
<1> I used temp tables and it worked beautifully
<6> no no, i meant something else :)
<1> for anyone who's interested -- this is literally the ONLY damned way to get independent randomization of both selects in a union that is then joined and randomized yet again
<1> oh, well what I did worked beautifully
<1> I am the champion, my friends
<1> hehe
<6> </sacrasm>
<6> ...
<1> pastebot
<1> http://sql-servers.com/nopaste/?show=335
<1> that was one bitch of a task
<1> thank god that's done
<1> ok, heading to bed
<1> g'nite all
<7> hi all
<7> I have a sql server 2005 table with an identity primary key int with a lot of holes between the primary key values. For example PK p*** directly from 234234 to 355432. Can i rebuild the PK?
<8> why do you want to?
<7> Because I usually make select using this pk with between clause
<9> omfg, why?
<8> so? you will still get them even if there's holes there
<8> BETWEEN 2 AND 10
<8> even if 2 and 10 are the only one existing in that range, you'll get those
<7> yes, using between 100000 and 200000 returns 90000 records, but from 200000 and 300000 returns 1 record only
<7> and so on
<8> so?
<8> if you want a sequential key then you need to either rebuild after each delete, fill in the gaps when you insert, or not delete at all
Return to
#sql
or
Go to some related
logs:
bug spincer
#flash
linkd network share
the most important province in canada
#beginner
irc from a nameserver
#windows
#redhat
#nhl
oxycottin heroin
|
|