| |
| |
| |
|
Page: 1 2 3 4 5
Comments:
<0> Arrakis: DBA
<0> not programmer... we just made offers to two programmers, so that **** is covered
<1> dba for hire?
<0> no, looking for one
<0> but you'd have to move to NYC (or the area)
<1> telecommute?
<1> feh
<0> feh?
<2> slightly more extreme version of 'meh'
<1> how much is the salary?
<2> contract or permie role?
<0> perm
<0> salary is probably pretty good
<0> they met my demands anyway :-)
<0> on the flip side, my bonus check still hasn't arrived :-(
<3> heh i blew up last friday when i got a $1k expense check and was waiting for a $6k one
<3> was ready to expense interest
<3> but it was rectified
<0> damn accounting/payroll weenies
<4> yooo
<2> what?
<0> huh?
<2> meh
<5> ok
<5> rmah on the hire again?
<5> what happened to your perl
<4> darken
<4> what did ur doc say
<6> hi
<6> to make it simple.. i do like this, I have two table now Sale - (attributes) - SaleID (unique), CustID, Date, SaleItem
<6> SaleItem -(attributes) - SaleID, ISBN, Qty
<6> for each, each customer can buy more than 1 item on a particular book
<6> I'm having difficulties in setting up a tracking pair book table now
<0> (attributes)?
<0> oh, the columns
<6> i have a pair table.. which i put PAIR ID(unique), ISBN A, ISBN B
<0> you need three tables: customers, books, sales (and ***orted payments, etc.)
<0> what's this "pair table"?
<6> let say if customer buy BOOK A n BOOK B, it'll be recorded into pair table
<6> if customer buys bookA n bookC.. it'll be recorded in pair table as well
<0> that's not the way you want to do affinity
<0> good god man
<0> that's insane
<6> so
<7> what if he buys books a b and c
<7> then youre ****ed :P
<0> what if he buys, a, b, c, d, e, and f
<6> mm.. so what do i do now?
<0> you think before you code
<0> how do you want to represent purchasing patterns
<0> how will time factor into it
<0> what about other factors such as demographics?
<0> will you be violating any patents?
<0> etc
<7> i think a good solution is to make a table with 1000 columns :)
<6> if i buy book a, the database has to list all the books which other customers has bought together with book a
<0> upb: haha
<7> each row is one order :PPPP
<0> Whoals: we all know what you're trying to do: amazon-style "suggestions"
<0> it's not as simple as it seems at first glance
<6> mm..
<0> I hope you're getting paid at *least* $100k for this
<0> bare bones minimum
<6> NO..
<0> haha
<0> kids
<7> yea i guess if you did some OrderItems table, connecting orders and books then the suggestion query would be very slow :P
<0> Whoals: a few hints: keeping stats on all book pairs is not a good idea
<0> what you want to do is keep a list of "favorites" for each book
<0> and dynamically update/prune them as orders come through the system
<0> sort of like a "cache" (in fact you can use some of the algorithms from sophisticated caching techniques for this)
<6> well, i have another question, i think it's a simple one - Hey i got this: SELECT ISBNA FROM PAIR WHERE ISBNB = 233; but it will display all ISBNA pairs (example: 111,122,111,123,124) How can i make it to display only 111? Since 111 occurs 2 times for ISBNB=233
<0> Whoals: for the last time, you don't want to do that
<6> ahh..
<6> what if i insist of doing that... is there anyway to do it?
<0> is this homework or something?
<0> whoals, you'll run out of disk space long before you have any useful stats
<7> how did youcome up with this pair thing anwaym its si weird
<0> consider: 100,000 books ... 2^100000 possible pairs
<6> learning SQL now.. come across with this question
<0> upb: inexperience
<3> Whoals: look into having and count(*)
<0> Whoals: keeping track of purchase pairs is the wrong way to solve the problem
<3> it scares me to think that some of the people asking questions already have jobs that they've claimed to know this stuff
<0> man, if you worked for me and was this stubborn, I'd fire you by now
<3> esp. when they have an example that fits your company and you know there are a pool of consultants work on
<0> heh
<0> phargle: if you actually know what you're doing and are reasonably well spoken, you can make plenty of $$ and never be hurting for a job
<0> when I interview people for positions, it's ****ing scary sometimes
<3> heh
<0> phargle: have you interviewed people for programming positions before?
<3> no luckily, it seems simple but the manager here seems to go through lots of candidates
<3> well, simple in the sense that, if you know everything about the skillset of the person your looking for
<3> if you don't i guess you need to have some way of having the candidate convince you
<3> which seems daunting
<0> dude
<0> you have no idea then
<3> heheh
<0> first, 1/2 the people just flat out lie
<0> it only gets worse from there
<3> i might get roped in to ask a few sql questions, but i know sql inside and out (esp in relation to my company's needs)
<0> I just ask that any "SQL expert" know more than me
<0> as I consider my SQL knowledge to be "barely adequate"
<3> are you a programmer?
<0> sort of
<4> rmah is a talker :)
<0> I'm a "consultant"
<8> lol
<0> which means I don't do squat
<0> :-)
<3> :)
<0> but really, I don't think my expectations are unreasonable
<0> where is CoJoNEsXXX?
<9> who here is familiar with wireless? can I not setup a huge *** antenna for the entire building?
<0> sure
<0> but for what use? that is the question.
<3> ooh, i think 4500 sq ft. is the cutoff for wireless. sorry.
<0> hah
<0> if he's talking about 802.11a/b/g, why doesn't he say so?
<3> i think they're sending him out to radio shack and he needs to know what to buy
<0> heh
<0> there's also the question of why he's asking on #sql
<3> heh
<3> hrm, mightve asked this earlier but.. in mssql i keep losing a full text index catalog
<3> i dont drop it, but it disapears. I can't find anything in bol that indicates a fulltext index ages and expires or deletes itself
<3> it's back now, but im obsessing over finding out why it just "goes away"..
<3> i see nothing in the logs in the timeline between 1 day where the fulltext catalog rebuild job works, and the next day when it fails
<3> .. i've resorted to doing auditing at the os level to see if some user is deleting it.. but that's just silliness.
<0> maybe you need a license key for it or something?
<3> hrm, not sure. i think the actual search mssearch.exe is part of windows
<3> but will check it out
<0> just a wild guess
<3> mm doesn't look it, but thks
<10> Can anyone help me out with the FreeText predicate
<10> for the freetext_string parameter of the FreeText predicate, i want to p*** in phrases and single words...The docs say if you want to perform a phrase search, enclose the freetext_string in double quotes....that works, but only for one phrase..i want to use a mixture of prahses and keywords....e.g. FreeText(Cloum, ' "My Phrase Here", keyword1')...
<3> i've only ever used just the phrase
<3> are you sure it accepts both (at once)
<3> er, or thats prolly what your asking.
<10> no i'mn not sure..but the docs don't say anything about mixing..
<3> egh, kinda drawing a blank but isn't one using the containstext predicate and the other is freetext or is that something else
<10> Maybe I should be using the contains predicate
<10> the docs say the CONTAINS predicate can search for a phrase or a word..and a lot of other stuff
<10> word | " phrase "
<10> <weighted_term> is what i want i believe
<10> Actually..I can use CONTAINS predicate..and the simple_term....e.g.: CONTAINS(Column, ' "My Phrase" OR "Keyword" ')
<11> I have a list which i'm trying to sort by streetname, streetnum, however the values are coming out incorrectly.
<11> the numbers seem to sort correctly, but the last number should be at the top.
<11> http://sql-servers.com/nopaste/?show=182
<11> please have a look
<12> StreetNum is not a numeric datatype
Return to
#sql
or
Go to some related
logs:
tasry
#stocks
#debian
#politics
#microsoft
blaekmon
young pornnet
fstab rehash
#ubuntu
#nhl
|
|