| |
| |
| |
|
Page: 1 2 3
Comments:
<0> DB2 has alot of child diseases
<0> on z/OS it's not recomended to use ORs but INs, when I did that in a join on LUW it denied to use an index, hence I had to use ORs ... silly optimizer
<0> and in fixpack 10 (pretty high) I couldn't update with a 0 ... they used several weeks to confirm it was a bug and deliver a fixpack 11
<0> do you have such problems in mssql as well ?
<1> God, no. :)
<2> "child diseases"?
<0> well, we call it that in norwegian
<0> when you're a kid you get ill very often like unstable
<0> we've moved to the latest v. 8.2
<0> hence the term "child diseases" :)
<2> Worst MSSQL bug I've ever run into was sporadic problems with parallelism with certain queries that made fairly extensive use of subqueries which was fixed in SP4. It's funny for years I never saw it once, and then all of a sudden I kept running into it.
<2> Fortunately there was a workaround; disable parallelism.
<0> hehe
<0> I hate that when old bugs comes back in new releases
<0> you wonder how they test the product before they deploy
<2> It didn't come back, I just never ran into it until after SP4 was released at which point applying the SP fixed it.
<0> k
<2> Bleh, tasked with writing a general purpose search tool for our solution and half of the data to be searched ain't normalized for ****.
<0> index it and you're done :P
<0> we're implementing fast for our intranet
<2> Not really indexable. It's garbage.
<2> Phone numbers, some with hyphens, some without, etc.
<3> so clean it first
<2> I guess we could put in place a nightly process to strip out punctuation from that crap.
<0> nightly ?
<0> you do that in seconds
<3> new crap added every day?
<2> New crap added constantly.
<2> Couple dozen times a day
<3> then first fix the front-end
<2> I'd love to rewrite all of the ****ty apps.
<0> I tested some replace functions on 40GB files and replaced 2-300k occurances within 1 sec
<3> how does the **** get into the db? stored proc?
<2> Mostly through INSERT statements
<3> ack
<2> There are probably 150 individual programs that do this ****, too.
<3> ACK
<2> I managed to spurn the project to create a single point of entry for I/O for our product, but we have a ****load of legacy.
<2> And a lot of it is GIGO
<3> indeed
<4> Hi can any1 please help with the following code http://sql-servers.com/nopaste/?show=369 , all i want is the results of that query in a new table called dupresources. Any help appreciated
<4> btw using postgre
<5> gosh, quitters
<6> can anyone help me with a solution similar to ebay's "pre-filled item information" i have a 6 table implementation but im trying to simplify it as best i can - no chosen rdbms yet
<7> Question: ms sql server. I have a table with the pk as a uniqueidentifier. I have the dfeault value set to newid() .. does that make it auto gen a val for the column when inserting?
<8> suicidel yes
<7> hrm. my stored proc says # params don't match table def.. but i have all of them minus the GUID
<7> hrn i added column names in my stored proc and it worked
<9> (sql2k ent) I'd like to p*** a CSV string to a sproc and do a SELECT * FROM Articles WHERE Title in (@CSVString) .. my CSVSTring is basically 'title1','title2' .. but it's not working.. any help is appreciated
<3> DevCodex : you can't do that
<9> yeah i figured that out .. i found a UDF that converts a CSV string to a table var
<9> any more efficient way to do it?
<3> nope
<10> fulltext index
<3> phargle : nope
<11> <11> How can I make sure that a new entry is created in a related table when an entry is created in another table? using sql server 2005. I'm looknig at triggers and ref. integrity but not seeing how to do this
<11> <11> for example TABLE B depends on TABLE A and I want TABLE B FK created when an ENTRY in TABLE A is created and initialized to default values. Id prefer not have to do this in code
<11> <11> I have the cascade option on but it doesnt insert new entries if they exist -- thats what I need
<11> <11> if they dont exist
<11> sql server 2005
<3> wLight : only do updates & inserts via a stored proc
<11> yuppie: Can you clarify and explain your reasoning and how it relates to my question? I ***ume you mean don't do updates from codebase (c#, C++, etc) directly to database but instead use the stored procedure?
<12> XSL/XPATH question: What is the syntax to add a predicate to the context node? ".[name(current())='Foo']" isn't valid but should demonstrate what I'm trying to do.
<3> wLight : that's exactly what i mean - you should never write directly to a table
<3> Sabba : did the title of the channel confuse you?
<3> Sabba : this is not #xml
<11> yuppie: Wouldn't stored procedures though lock me into using a particular database vendor versus using generic provider-indepedent client code (ADO.NET abstraction)?
<3> ffs
<3> why didn't you say that in the first ****ing place
<3> i hate people who only state half the problem space
<3> then whine when you give them the best answer to what they *have* stated
<3> so **** off
<3> you too
<10> YuppieScm: actually, it would.. have you ever used one?
<3> yes, but i fail to see how fulltext index will be a better - or indeed any - solution for the expressed problem
<10> i find a lot of ignorance surrounding fulltext indexes..
<10> im not saying it's the best, or more efficient than putting something into a table variable.. but if you know about them it can be a valuable tool.
<3> agreed it can be a valuable tool, but
<3> i fail to see how fulltext index will be a better - or indeed any - solution for the expressed problem
<3> anyway - i officially don't care now
<10> you can p*** the string intact
<3> i'm off to egypt for a holiday
<10> heh, hence my ignorance remark.
<3> p***ing the string intact is in fact what he's going to do
<3> then parse it into a table variable
<3> anyway
<3> good bye for a week and a half
<3> unless the hotel has wifi
<13> Hi, im uring an MS sql-server database and have the following issue:
<13> i have a table with ContactId's and with GroupId's. The contactId will occur often depending in how manny groups he is in. I'm not trying to build a query that shows me with a boolean in what groups the contact resides
<13> -not
<13> (i AM trying to build the query)
<14> problart a stupid f* question .. but .. http://sql-servers.com/nopaste/?show=362
<14> thing is, i've seen references from subselects like that before .. why the hell it wont pick up on "chain c1" from the subselect is beyond me :(
<15> pluisje: with a boolean?
<15> you mean a rs like ContactId Group1 Group2 Group8 Group123 ?
<13> yeah
<13> exactly
<13> is that even possible?
<15> yep with dynamic sql
<13> is it possible in 1 statement?
<15> sure 1 statement but you have to generate the statement
<13> yeah, i figured
<13> but do i have to use ali***es?
<15> yeah
<13> is there some tutorial for this?
<15> you can do it in oracle like: SELECT c.ContactId, SUM(DECODE(c_to_g.ContactId, 123, 1, 0)) as Group123, SUM(DECODE(c_to_g.ContactId, 888, 1,0)) as Group888 FROM (SELECT ContactId FROM table GROUP BY ContactId) c, table c_to_g WHERE c.ContactId = c_to_g.ContactId
<15> smth like this
<13> i'll try it in sql_server
<15> no it wont work ;)
<15> you have to adapt it
<15> mssql doesnt have decode i believe
<13> Msg 195, Level 15, State 10, Line 2
<13> 'DECODE' is not a recognized function name.
<13> what does decode do?
<15> compares the first argument to second, if its equal evaluates to third argument
<15> then compares to fourth, if its equal evaluates to fifth etc
<16> Hello all
<15> otherwise it evaluates to the last (Default)
<13> hmmz
<16> Trying to make an SQL query, with a where clause which has a condition on a field (called length) of type interval - the length is for example.. 01:15, 02:29 (hours:minutes)
<16> how do i select all with length > 2 hrs ?
<17> Peach44, state your rdbms
<16> postgresql
<16> SELECT * FROM blah HAVING sum(length) < interval '2 hours';
<16> thats what i'm doing at the moment
<16> giving me error: Attribute blah.attrib must be GROUPed or used in an aggregate function
<17> that doesn't look valid
<17> right :)
<17> why are you summing up ?
<16> because thats what google told me to do :p
<16> tried to look it up... SQL queries regarding attributes of type interval are scarce
<17> it's just that 1:15 that has to be higher than 2h right?
<16> yes
<16> so greater that 02:00
<17> well I don't know postgresql
<16> what would SQL be for it as a starting point... ?
<17> can't you just do "where length > '2:00' " ?
<16> wonderful
<16> thanks :)
<18> Hi. :) Im trying to delete a row in my MSSQL database and i've written a stored procedure: DELETE [*] FROM table | This doesn't work.. Why is that?
<15> hmm, whats [*] ?
<15> DELETE FROM table WHERE id = whatever
<18> I wanna delete the whole thing...
<18> Not just by id's
<15> you siad you want to delete a row?!
<18> Well, i do... Hehe..
<15> DELETE FROM table WHERE id = whatever
<18> Im kinda new at this, sorry...
<15> this deletes the row where id = whatever
Return to
#sql
or
Go to some related
logs:
#photoshop
#nintendo
zeritex asthma
#redhat
#heroin
#politics
#worldcup
ignore user from on TEXT +mirc
Billian rapidshare
Swedish dish system
|
|