« first day (385 days earlier)      last day (4497 days later) » 

4:40 AM
@NickChammas when you're about tomorrow ... since I'm headed to bed
should both show as unregistered when you goto their page. I can't do anything about those, until one is registered. You should be able to see that. The corrective action in the future is to ask them to register one or the other account. I've got to refresh my memory how that is done (create an OpenID? actively login? I forget) but for now, goodnight all
 
 
1 hour later…
5:52 AM
@jcolebrand ah ok; I didn't check
 
 
2 hours later…
7:34 AM
Hi. I have a primary key int field on one table in sql server...can I alter that column and make it an identity column?
 
 
4 hours later…
gbn
11:08 AM
@NickChammas of The Sun?(NSFW)
 
 
1 hour later…
12:09 PM
@JNK That's because it doesn't have to do a sort isn't it?
@JackDouglas SQL Server does not implement any facility to seek directly to the next distinct value in an index skipping duplicates along the way. I thought that was what the skip scan did in Oracle. Guess I was wrong about that then? For SQL Server if there are lots of duplicates then a recursive CTE can be faster (repeated index seeks) as @Sqlkiwi illustrates on this post sqlservercentral.com/Forums/FindPost1013407.aspx
 
 
2 hours later…
JNK
1:48 PM
@MartinSmith I think it's sort-related, yes. the point is having an index on the appropriate keys does make a positive impact.
 
2:20 PM
@JNK re: dba.stackexchange.com/a/11459/2374 Cost percentages in a plan are always "estimated costs". There is no relation to the resources incurred.
 
JNK
@MarkStoreySmith - yep I know they are all estimates. However there is a unit based on some combination of resource usage (estimates).
 
@JNK My suggestions is to clarify that the costs are estimated, rather than being an indication of the actual costs incurred when running the query and seeing an "actual plan". Your answer reads (to me) that the percentages are a reflection of the actual costs incurred to run the query. If you see what I mean :)
 
JNK
Gotcha, I'll take a look.
take a gander at the revision and make sure it's clear
 
@JNK All good!
 
JNK
thanks for catching that
 
2:45 PM
Good news! The vendor database removed all default values so none of our insert statements work :D
I think these guys aggressively hate data integrity
 
3:23 PM
Is there like a best-practice way to add an index on a huuuuge table in production? I need to add an index that takes half an hour and at some point I need to move it into production. Best thing I can think of would be to schedule adding it on the weekend when the system isn't in use, that's when we run backups and such...
 
3:41 PM
Crisis averted, the index doesn't help anyway. Which brings me to a better question...how the hell can a SELECT query get 4x slower after adding a non-clustered index? I thought only DML queries got slower
 
@BenBrocka If the optimiser thinks it will be faster to use the index... but it isn't
Does the execution plan show it being used?
 
Eff, I already deleted the index and closed the window with that execution plan. It was an index suggested by SSMS. It obviously wasn't helpful. The execution plan w/ the suggested index had much fewer steps but I didn't notice what it was doing
I'm curious but don't really want to wait another 30 minutes to generate the index again
 
4:07 PM
@MartinSmith a skip scan does something like the opposite - matching on non-leading fields. Thanks for the link - Paul's post does something similar to what I do with a subquery in the select - force multiple 1-row range scans. Seems like a fairly cheap optimization for the DB to save the hassle though I wonder why they haven't implemented something like it internally.
 
gbn
@BenBrocka do you introduce a key lookup? That is, the NC index is not coverng?
 
@gbn No, but I hear there are some fun WTFs in that paper too.
 
Should be covering I think...between ON and INCLUDE it covers all relevant columns
 
4:31 PM
@jcolebrand : hi. why u ping me?
 
@BenBrocka Maybe there is some correlation with your statistics that does not fit the optimizer's assumptions, for an example see
6
Q: SQL why is SELECT COUNT(*) , MIN(col), MAX(col) faster then SELECT MIN(col), MAX(col)

CodingBarfieldWe're seeing a huge difference between these queries. The slow query SELECT MIN(col) AS Firstdate, MAX(col) AS Lastdate FROM table WHERE status = 'OK' AND fk = 4193 Table 'table'. Scan count 2, logical reads 2458969, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical rea...

@JackDouglas Thanks for the clarification on the skip scan.
Although actually the issue on that one was key lookups. If you want to generate the execution plan without actually building the index you can do some experimentation with hypothetical indexes and DBCC AUTOPILOT
 
4:47 PM
@MartinSmith Unfortunately I have other stuff to work on right now. Interesting though.
Thanks
 
5:15 PM
@AlexMathew I have no clue. Was it in here? Let's review the backlog
Jan 19 at 21:04, by Nick Chammas
what would be appropriate protocol? ask if he's interested in funneling people here?
 
5:34 PM
@AlexMathew - We noticed you are the owner of the MySQL chat room on SO and wanted to let you know we have a lot of database dev and admin activity in this room. Both you and the regulars of that room are welcome to hang out here.
 
And so the DBA vampires set upon their latest victim...
Before long, he would be one of Them!
 
@BenBrocka hehe, yes we are :D
 
5:53 PM
@jcolebrand do you know how the mod pro temp "elections" work out for new sites?
 
@BenBrocka yes, and you should too sir :p
it's outlined in the 7 meta Q
basically, you need to have a post nominating yourselves as a CW
one person, one answer
then the answers can be upvoted
then you approach rchern or someone and say "please make us pro-tems" but they'll do that about 3-4 weeks in
once you go pubbeta, that's when you need the meta Q
I wouldn't make it during privbeta, you need to sort out that loose cannon and delete more lousy Q
 
Was thinking it was something we do at the start of public beta (so 3-4 weeks is the length of private beta?)
 
nah, a privbeta is as long as it needs to be to show a concise vision of what the site is about
pubbeta is as long as it takes to prove the site will have people coming back
 
Ah. Currently I think we (cog sci) need more, better questions per day, activity has dropped off. We've dealt with most of the worst of the "spammed" questions...
 
Aye
now you need to get good content from reputable sources
Josh is one of my buddies, we skype a lot
so I know all about cogsci ;-)
 
 
1 hour later…
7:03 PM
Unfortunately we don't seem to have all that many experts or practitioners in the field
 
0
Q: Data uploading from excel to database

samarI have one excel file which is updated daily, and I have to update the records in database(sql server 2005) daily. What will be the query to fetch data from excel and update on database? e.g the table structure like: Empid EmpName Emp.Des 001 Robert Engineer 002 Philip Trainer 003 J...

this is pretty basic isn't it?
 
@jcolebrand Yes. It could live on SO or B.I. enabled dba.se equally well but probably not much to be gained from migrating it.
 
@jcolebrand Do you have any experience with private betas?
 
8:07 PM
needs one more vote to make community ad
 
8:21 PM
@NickChammas done
 
8:40 PM
@SimonRigharts Bam. Ad now shows up on main site. Sank yuu.
 
@NickChammas Allegedly UX has had success with that style of ad
 
@BenBrocka success, as in lots of twitter followers?
 
@NickChammas as in a fair amount of followers considering the size of the community
We discussed it's CTR and it seemed sort of tepid but I have no idea what to compare the CTR to, it's got an okay number of followers for a smaller site though, 450
 
CTR?
 
Twice as many as @stackDBAs though I';m not sure how level the playing ground is
Click Through Rate, people that clicked the ad divided by total impressions
 
8:55 PM
@BenBrocka well, our ad just started showing 5 min ago :)
if that's a significant source of followers, then...
 
Problem is we don't have a method to track conversion rate; #/% of people who actually clicked through AND followed because of the ad
 
-1
Q: Attach database failed for server localhost\sqlexpress2009 (Microsoft.SqlServer.Smo)

user1161815Getting access denied when trying to attach AP.mdf. I can see the files at the bottom of the screen. Error is: Unable to open the physical file "C:\Murach\SQL Server 2008\Database\AP.mdg". Operating system error 5: "5(Access is denied.)". Microsoft SQl Server, Error 5120)

@jcolebrand @gbn @MarkStoreySmith VtC as too localized.
@BenBrocka yah, that would require some sort of feed from twitter
 
@NickChammas I think you might be able to track that if you use a "follow" button directly but I don't know if we can make a link do that
How long has DBA been around, counting public beta?
 
13 months?
a whole bunch of people got yearling badges this week
all of them members for 1 year
so probably 12-13 months
those people also have the Beta badge
 
Hm, we've been around longer and have 40% more users, so I'm not sure I could attribute the difference to the ad
Er, not 40%, like 60%. Some number
 
9:13 PM
Is it worth posting questions if I'm almost certain I won't get an answer?
Thinking about putting one up about sub-optimal query plan selection in Sybase
 
@SimonRigharts if it's an appropriate question then fire away
 
Depends why you don't think you'll get an answer. But if it's a real, on topic question I would post it and let the community figure out
You'll never know, nor will we, until you ask it :)
 
you can then share the post with anybody who might be able to help
potentially recruiting a new Sybase DBA.SE person
 
That's the annoying part - the only person I know who might be able to help charges $150/hour :v
(the reason I don't think I'll get an answer is because I seem to be the only person answering Sybase questions, and the question I have in mind is along the lines of forcing specific operators in a query plan)
 
does sybase have query hints?
in any case, we might be able to give the general advice, and you'll only need to look up the syntax
 
9:21 PM
yes it does, up to a point
the other thing is that, since it's a performance issue, the problem doesn't really manifest until you get to a non-trivial data set size
(the problem tables in this query are 90 million rows, 60 million rows and 100 million rows, collectively weighing in at 60GB of data)
 
@SimonRigharts Do you have a mirror or prod support box you could test it on?
 
@NickChammas We could really (apparently) use some of those. I don't know of any other users that know it
 
well we have like 30 copies of this database all up, so yeah I have test boxes coming out my eyeballs
2
 
Maybe you could frig the indexes and hints on one of the test environments.
 
10:12 PM
@NickChammas I am tempted to leave it open as a valid troubleshooting Q. I'm just one man, however, and not perfect.
@BenBrocka I participated in this one? I've been on metaso for ages? I was in a couple other that I couldn't commit to as well as I liked?
 
10:38 PM
@jcolebrand I see. I'm just trying to think How I can help the community along in the beta. I've done lots of community moderating clean up stuff but I don't know how I can help us get people more active
 
@BenBrocka get out of privbeta?
or did it go to pubbeta?
 
We're still in private beta, just a week old now
 
@jcolebrand okie doke
@jcolebrand Why are these all questions?
:P
 
Maybe he's a teenage girl? And he has inflection like a question? To the end of all his statements?
 
10:53 PM
LOL
yes, that's what I was getting at
looking at your profile @BenBrocka, you must know a few too :)
 
Not as many since I gradumatated. And most people where I work are over 40...it's weird
 
yah, that was my first job too
big financial services company
people there had kids my age
 
> Children with this condition often suffer oral cavity damage both in and around the oral cavity (such as having bitten off the tip of their tongue)
Oh the things I learn from the Cognitive Sciences site
 
11:44 PM
Was my comment on this question too harsh? dba.stackexchange.com/a/11500/3761
if he'd spent a hot two seconds looking at the link already provided he would've seen what he's looking for
 

« first day (385 days earlier)      last day (4497 days later) »