| |
| |
| |
|
Page: 1 2 3 4 5 6
Comments:
<0> hehe
<0> dfworking, 1 year ago i tried somth enclosed by brackets and i was able to get results
<0> at that time i did some research about wildcards ..
<0> mmm
<1> not sure what to tell you
<0> dfworking, [C-F] works for example
<0> displays all the courses starting with C and F
<1> thats a standard LIKE wildcard
<1> yes
<0> i want it to display all the courses starting with CSC and FAP
<1> sec, not sure if thats doable
<0> [CF] returns the same result as [C-F]
<0> mmmm
<1> guess you can try [CF][SA][CP] but you run the risk of getting combos you don't want
<1> such as CAP, etc
<0> that doesn't work
<2> how do i select the first 100 rows in oracle?
<2> found it out: where rownum <= 100
<3> This is probably easy, I just can't think. I had a brilliant user delete rows in an Access database and I'm trying to basically merge an older version of the same table. I want to create a query of all of the rows (based on the ID number) that are missing from the old one so that I can do an INSERT on that query once I've got it right. Is it doable or do I have to get into programming?
<3> (that's not an issue, I just wanted to keep it all in Access if I could)
<2> on oracle: I am trying to empty a log table with "delete from logentry", but it seems to never finish.. It has been hanging for 5 minutes now, there are about 5 million rows in the table. Are there any faster way of doing this?
<4> does truncate work in oracle?
<5> truncate table logentry;
<6> bealtine: I think truncate is an ansi standard ... for -92
<4> probably but I'm too lasy to look it up for him :)
<2> cool, thanks!
<7> what's the best way to deny access to the base tables in mssql? run an sp_msforeachtable denying access to public?
<5> JohnD: truncate isn't exactly the same as delete. You should read up on it in the bol just so you don't use it when you shouldn't.
<2> yeah i found out it doesn't log the deletes
<2> anyway i don't think i can delete all rows, i have to use a where clause that just deletes 95%.. So it will take a while. strange
<2> oh the joys of using oracle
<5> That's not an oracle thing. :)
<2> well i have never seen a delete from taking over 5 minutes on other databases
<5> Sounds like indexing problems, or you're not limited on columns with useful indexes
<5> Or both.
<2> if i do "delete from myTable" (without a where clause) indexes shouldn't be a problem?
<5> Proably not, but a logged delete of 5 million rows is pretty i/o intensive. I'd look at the configuration of the server machine at that point.
<2> okey thanks for helping
<2> another issue we have with oracle is when rolling back transactions, it doesn't roll back create and drop table statements.. Is this a known issue or something we are doing wrong?
<5> I don't think oracle logs DDL statements.
<5> Its pretty random who does and who doesn't, and changes from rdbms to rdbms.
<2> okey
<5> Oh, wait. Even better.
<5> Don't use DDL in the middle of a transaction. Its bad in Oracle.
<5> And this is just weird. :)
<5> "We know a DDL (data definition language) SQL statement implicitly commits your transaction, even if this DDL fails (implicit means 'user commits' statistic does not increment)."
<5> http://www.stormloader.com/yonghuang/computer/OracleIdiosyncrasies.html
<5> :)
<8> That's weird
<9> it's reverse psychology
<10> Hi, I'm using MSSQL2000. The need has arisen for a "working table" that will constantly get populated with 1000s or rows, and as frequently, it will have 1000s of rows deleted from it. These additions and deletions will ultimately be invoked programatically (.NET SQL Library or similar), but on the database level, I'm wondering what the most efficient way to do this is. I'd like to avoid row-by-row inserts, and BULK INSERT...
<10> appears to only work with a file. Thanks for any thoughts!
<11> ehm, what sort of solution are you hoping for? you dont wanna do it row by row, and yet youre not willig to import from a file ?
<2> maybe osmosis?
<10> i thought maybe there was a programmatic BULK INSERT of an dataset or ADO Recordset...
<10> trying to avoid a lot of flippity flopping on the program side to prepare it for the insert.
<8> Using .NET 2.0?
<10> .NET 1.1
<8> 2.0 has the SqlBulkCopy cl***
<8> There may be third party libs to access that pre 2.0
<10> hm. maybe ill see if i can use 2.0 then. i'll google around for SQLBulkCopy
<8> Another route you can take is to format the data into XML and use OPENXML to translate that into tables.
<10> thx.
<11> why not row by row ?
<10> yeah, i tried that but it is too expensive
<11> doesnt mean you gotta fire off 1000 queries
<8> Send it in batches.
<11> just send them as one big batch
<10> row by row wouldn't be ideal, according to the DBA, because of the expense
<11> or yeah, in chunks
<10> pardon my ignorance
<10> but what do you mean by batches?
<11> well you can run 50 updates, but theyre all in one batch
<8> Well, sending up batches of INSERT statements requires dynamically constructing SQL.
<11> yeah true, not ideal either, just an idea
<8> Which works fine, but scares some people particularly with the level of data validation required to prevent injection.
<10> thats cool. ill ponder batching too.
<8> XML is nice because the data is preformatted and you can easily translate out into tables including heirarchies.
<10> i know SQL2005 has all kinds of cool stuff it can do with XML
<10> so maybe XML is the path forward even now
<8> SqlBulkCopy cl*** makes use of an undocumented API in the TDS protocol which allows the client to send data up in a binary block of data which is cached server side and then p*** an INSERT BULK statement which translates that into a table.
<12> hrm
<12> wake up you stupid ****ing bot
<12> ****ing mirc defaults ****s, /m should be aliased to /msg by default ...
<13> MoZ. stfu kthx
<13> fag!
<14> What data type should I make a field for MSSQL if I want to store Japanese characters in it?
<15> anything thag begins with an n probably
<14> thanks - it seems to be working with just varchar but I wondered about the nvarchar
<16> Hey, what is sql?
<5> ...
<14> phil_MR : Don't you think you could find a good answer to that on Google? :-)
<16> my mom turned off searching
<16> so i don't look at pictures
<5> Not even worth the time. :)
<10> thanks halo. that is sounding very good, after a powwow with the DBAs
<4> illustrated guide to sql ?
<4> by dilbert
<14> phil_MR : SQL is Server Query Language... It's a database system.
<4> structured query language...
<12> aside from dataset and datatable ... what are some other 'recordset' like objects in ado.net ?
<14> phil_MR : Does your mother also walk you on a leash when you go out of the house?
<12> I'm trying to find something that's similiar to ADODB.RecordSet ... but supports multiple result sets from a stored procedure
<4> datareader
<12> datareader requires you to know the columns datatype before retrieving it ... no?
<4> ja iirc
<12> yeah, thats counter productive ... ';(
<17> anyone konw of a channel that's dedicated to sharepoint portal service?
<17> I know this isn't a pertaining channel, but figured someone might know.
<12> dataset supports more then one result set ... or well, it can work with multiple tables ...
<12> if I fill it with the adapter, dataset.Tables.Count is the number of results returned. ..
<12> just can't figure out how to get to the damn results, haha ...
<18> samalex i don't think so; i've been looking for it too
<17> there used to be one on efnet a year or so ago, but guess it dried up.
<19> is there a way to sort by 2 criteria?
<19> like SELECT TOP 5 * FROM tbale1 ORDER BY date, type desc?
<19> or SELECT TOP 5 * FROM tbale1 ORDER BY date, table2.type desc?
<3> Gm0nEy: SELECT * FROM table1 ORDER BY date ASC, type DESC LIMIT 5;
<19> even if type is from another table
<19> ?
<3> Gm0nEy: SELECT * FROM table1, table2 WHERE table1.someID = table2.someID ORDER BY table1.date ASC, table2.type DESC LIMIT 5;
<19> ty
<12> Quant ... you there?
<19> Weezey, if i want to grab the last 5 records(by date) that have a specific type (table2.type)?
<19> do i set @type =1
<3> sort it backwards, then do limit 5
<8> LIMIT isn't valid T-SQL
<8> But the premise is the same, the order by applies prior to the record counting
<20> yea, I wondered what DB Gm0nEy was using
<21> Is there a place where i can tell MSSQL what data format it should use when the date format is this: 20060209:07:28:25:703. I have it in a varchar attribute at the moment and want to alter the attribute to a date
<19> im useing SQL2K
<22> SilverWol that doesnt look like a valid date format for mssql
<19> but the problem is this
<11> :)
<21> SouLShocK, can i in someway make it valid?
<11> format it yourself using SUBSTRING
<22> format it
<11> '2005/01/31 09:55'
<22> or 2005-01-31 09:35
<19> i have a two tables one table has is written to when new docs are put up and ***igned a key
<19> then that key has a relationship to another tables key and there the same docs(by key) are tagged in a column by type
<21> Carpathia, oki Thanks
<19> so in the first table i have the docs dated
<21> SouLShocK, Thanks :)
<19> so iwant to sort by date and type
<19> so only latest type1 docs, maybe my db design is bad?
<19> i gotZ it
<19> SELECT TOP 5 * FROM tblTITLES, tblPUBTYPES WHERE tblPUBTYPES.type = 1 ORDER BY tblPUBTYPES.type DESC, tblTITLES.date DESC
<7> got an interesting mssql security problem
Return to
#sql
or
Go to some related
logs:
cannot join * (Requires keyword)
#stocks
#windows
mirc israelpass
#beginner
hkgolden video card
#beginner
#politics
#gentoo
#beginner
|
|