| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8
Comments:
<0> how much can text you put in a text datatype in ms sql, I see the length is 16
<1> i think its like 2-4 gigs or something like that
<2> 2gb
<1> i guess the value of an int? 2,147.....
<2> 2gb
<0> ok thanks, why does the Length say 16 in sql 2000?
<0> when you create a column of that type
<1> I think its a pointer
<1> text data is stored on separate pages
<2> because the data for a text field is stored seperately from the rest of the table
<2> for performance reasons
<2> f209 : you know, all of this information is in BOL - why not try hitting F1 once in a while
<3> how can I format a numeric column to return a value of 5 as 0005
<2> Narbs : in the presentation layer
<2> Narbs : sql is not the place to format data
<3> oh ok
<2> Narbs : which is why sql has no formatting functions
<1> You could use a trick like this though: SELECT RIGHT('00000' + RTRIM(cast(column as varchar(5)), 5) FROM Table
<2> Eol : you could, but performance blows, and it's just the wrong thing to do
<2> there are any number of ways to do it, but none of them are right
<1> Yeah, but sometimes you don't care about "right" and just need it done
<2> Eol : first you teach the right way. when the lesson is learned, only then do you discuss the wrong ways
<2> Eol : the right way is not in sql
<2> Eol : how about we try not to teach n00bs any bad habits?
<1> Who's to say he's a n00b or not?
<2> Eol : i am - because he's asking a n00b question
<2> Eol : ergo, for the sake of this argument, he's a n00b
<2> Eol : n00b is as n00b does
<2> Eol : so, get a grip - the right answer is "not in sql"
<2> Eol : if you want to give people the wrong answer to question, why don't you start your own channel?
<1> I didn't give him a wrong answer, my answer, does in fact, work
<1> It is not the "best practices" answer
<3> I'm trying to change a data type in sql server management studio on a large table and the operation is timing out. any way to change the timeout period?
<4> the real question is.. why are you trying to change a datatype? have you researched the possible ramificaitons of it?
<3> char(20) to varchar(20)
<3> it was char because the table was being populated from an oracle db that used fixed char types
<3> but that's not the case anymore
<5> script the action and run it manually.
<6> stupid netsplit...
<7> indeed
<8> How can I return unique results in a full text search using CONTAINS
<9> sounds like homework
<10> how to insert multiple (2,000ish) rows into oracle with one line
<8> nope not hw
<11> terminus, use sqlldr or external tables
<11> or turn off wordwrap and make it a LONG line
<11> k00, use the distinct keyword. who would have thunk it
<9> how are you getting 2000ish records? that would dictate how to do it
<11> im thinking... excel
<10> the 2000 records are generated from my program that parses documents
<10> :-)
<10> the records are documents
<9> and what does your program do with them?
<10> parses and stems them
<9> and puts them into a file? or what?
<10> nope, stdout
<10> and then i use OCIExecute in a php script to insert them
<11> lovely...
<10> indeed
<10> it's a generalized vector space model search engine
<11> is it now
<9> uh huh
<11> not much of a search if you cant figure out how to get data into it
<10> getting data into isn't a problem
<11> seems like it is for you
<10> nope
<10> 1986 rows selected.
<9> so right now you are running single insert statements?
<10> doing it in one statement is the problem
<11> why do you need it in one line?
<10> i was using an insert all with select union alls
<10> because doing 2000 OCIExecutes is a pain in the butt
<10> :-)
<9> im not sure what tools php give you to work in bulk
<12> you mean it's slow
<10> it's not php i'm worried about
<10> it's the sql
<10> tenfour, yep you got it
<12> can you use XML?
<9> its probably the php connections
<11> no its the whole idea
<10> no **** ^ like i been saying :-) hence why I want to only do one OCIExecute ( aka one line)
<12> this doesnt sound appropriate at all though
<12> i mean, why is your app inserting 2000 items in a table?
<9> do you have the records in bulk?
<10> i have n documents i want to store in a database right? and I am parseing, and stemming the content of these documents
<10> and then i am storing the document #, the content, and the stemmed terms in the database
<11> oracle has all that built in
<10> yes it does
<10> it does not do what I want though
<9> what kind of document is this?
<11> you dont know how to do the very basic things
<10> does it matter?
<9> yes
<11> so I doubt you know if it does what you need it to do
<10> you know, it was a simple question
<13> anyone have a lot of experience exporting relationships using FOR XML AUTO?
<9> its not
<12> terminus your question is worded simply but it's really not a simple syntax question
<8> Jarett_, how would I incorp the distinct keyword in this piece of code: http://sql-servers.com/nopaste/?show=295
<10> well here are several answers that could have been given, which I have tried: PLSQL, a multitable insert using UNION ALL with selects, or as with mysql you could do insert values (x,y); insert values (a,z); <- doesn't work in oracle
<10> so yeah it is a simple syntax question in some respects heh
<11> sqlldr is the right answer
<11> another acceptable but not as efficient would be bulk binds and insert
<10> yes thanks
<11> "single line" is an improper thought
<10> i'm not too concerned about the efficiency of loading the documents as long as it's less than say 30 seconds
<10> because it's a one time thing then my queries are driven off an ad hoc database
<9> depending on what you are storing and connections and memory,etc 30 seconds sounds pretty arbitrary
<11> in 1 minute i could write a php parser to create a sqlldr control file, export the data to text or p*** it as a variable, and then exec out to sqlldr
<10> Jarett_, that is a great idea and I'll pursue that
<11> or if its 9i or 10g you just do it via external tables
<10> it's 10g and yes using external tables was what I was doing
<11> but if its coming from another data source, there is no reason not to connect directly to that data source and pull the data
<10> the data source isn't something you connect to it's documents on the web
<9> oracle can read from a website
<10> i'm grabbing the content from them and giving the docs a unique identifier
<11> well pull them in as clobs, parse them in the db, insert as you parse
<11> ummm easy
<9> I have a plsql proc that reads from a website parses and inserts
<11> given a text (like xml or webpage) i can parse out anything, and convert to almost any other document format.. or break it down
<8> hrm you can't select a distinct ntext field
<10> yeah what i was doing was grab doc-> parse -> remove stop words -> stem -> insert
<10> in a very simplified breakdown of the operation
<9> yeah you can do that in a stored proc
<9> UTL_HTTP might be a starting point for you in that case
<11> plus why do you need to remove stopwords?
<11> they aid in scoring and searching
<11> if you want to search only in <title></title> or in the body.. or in different CSS tagged area
<14> see http://sql-servers.com/nopaste/?show=293 ... using mssql, im trying to find an easier/more efficient way to rewrite this query. i feel like im joining my tables too much
<11> thats pretty ugly aLfer
<14> yeah i know
<14> how can i redo it?
<11> break your tables out of the 3 individual selects
<11> theres no need for it. either do 3 selects and union all.. or do 1 select with subselects
<14> isnt that what i have now? 1 select with subselects?
<11> no
<11> you have 3 selects wrapped in a different select which is totally pointless
<14> what is the syntax for a subselect?
<11> www.google.com
<8> how do you select distinct ntext rows
<14> i think the problem is i have multiple aggregate functions in the select line, and each one is based on different criteria.. so i have to use an individual select to be able to use the different where clauses
<11> learn to use group by
<13> so without resorting to FOR XML EXPLICIT, is there a way to use FOR XML AUTO to preserve and nest relationships in generated XML?
<13> I've read about 15 MSDN articles and they all use EXPLICIT to get the job done, but I don't need that level of detail in what the tables contain
<14> jarett_: okay i rewrote the query http://sql-servers.com/nopaste/?show=296
<14> now i get Column 'joined.prstart' is invalid in the select list because it is not contained in either an aggregate function or the GROUP BY clause... but the problem is i dont want to group by prstart, i want them summed
<12> why are you trying to return 3 completely different values in that query
<12> if you want 3 different kinds of aggregations, just do 3 queries (like you are already doing)
<12> it would also be possible to put it into 2 queries probably, by using with rollup
<14> well i wanted the three aggregations to be lined up for each project
Return to
#sql
or
Go to some related
logs:
#linux-noob
Flash scroll images endless loop
#beginner
#beginner
canon 30d training video torrent
#computers
streamtvnow shoutcast
Ted Strykar
urxvt LC_TYPE
spratt tutorial rhinoceros
|
|