« first day (841 days earlier)      last day (4019 days later) » 

12:00 AM
Seriously.
I am taking a break from these f*ckwits across the SO network. Totally done with it,
@Bill s = start, e = end. Pretty simple. But ok, use a recursive query that could cause performance problems. shrugAaron Bertrand 1 min ago
I'm sick of spending time crafting answers that are rejected for idiotic reasons.
 
 
1 hour later…
1:14 AM
@AaronBertrand It's okay.. i got serial downvoted... probably by JW for telling him to stop editing tags.
0
Q: since when am I allowed to cast delete votes on questions?

FreshPrinceOfSOI don't have 20k rep, yet today I noticed the delete option on a question. I thought the vote to delete privilege doesn't start until 20k?

 
 
3 hours later…
4:59 AM
An open bounty prevents close votes. Who knew!
2
 
 
3 hours later…
7:53 AM
wow what a great and welcome edit.
6
A: How to create a database for unknown kinds of data?

Chris TraversUnknown kinds of data sounds to me somewhat fishy. Your examples, of course, are all knowns. For goods and services careful analysis and normalization are important and I think you can get away from EAV modelling (which I think will cause mor problems than it solves) for core data. The rest co...

wait
wrong one
that was my answer
the other one, one sec
2
A: How to create a database for unknown kinds of data?

Paddy CarrollDont use MySQL, a relational database is not used for the solution of this type of problem. Use a document or NoSQL database such as MongoDB or possibly RavenDB on windows. Or alternatively use PostgreSQL. If you have a base set of attributes you can build inheritance into your tables create ta...

that's the right one
 
 
1 hour later…
8:57 AM
Morning all
 
good morning
 
9:09 AM
Morning
 
@ChrisTravers yes, I was pretty surprised and impressed with that
 
It's definitely a case for reversing downvotes (which it looks like a bunch of us already did)
 
gbn
Morning all
 
Morning
 
9:35 AM
Hello
 
9:45 AM
Some people go out of their way to create trouble :-P
Just to clarify that.... The only difference between synchronous_commit being on or off is that it waits for the standby to say it has committed before returning this to the client. If your answer to "the slave went down" is to turn replication off, you are getting nothing out of the additional guarantee that synchronous commit gives you, which is that once the commit is finished the data exists on at least x number of servers. You will have far fewer issues by simply not offering the guarantee in the first place than by offering it except when you don't. — Chris Travers 36 mins ago
(Doing streaming replication and synchronous commit with only one backup server gives you failure anytime either system goes down....)
 
@ChrisTravers Can you help me explain this diagram: apandre.files.wordpress.com/2011/05/cube-slice-blue.jpg
If you are wondering what it is.. then you probably can't help me :D
Well, I want to explain it myself first
 
I understand it. What are you trying to do?
 
Ok.. I want to understand how it works. Shall I go ahead and explain how I think it works?
 
Ok, the cube is basically a stacked 2d array of of matrixes. A row is represented by products, and the columns by continent. Which means a x, y cell is the measure that relates to a product for that country. However, the third dimension.. the z axis represents time
So the matrix is organized (or stacked) by the time dimension
The first 2d matrix gives me all products per continent for the year 2000, the next 2001 etc..
Is that far off the mark?
 
9:54 AM
Ok, so this is a logical design for what you can aggregate in a database.
(probably a data warehouse)
You want to think about it in terms of what slices you can take
these slices could be one dimensional or two dimensional
For example the highlighted portion gives you all sales of all product categories in Asia by year
does this make sense so far?
 
Yeah it does, finally I think.
But I could have a horizontal slice
 
ok... so typically this would be accomplished by what's called a star schema.
 
That doesn't involve time?
 
right
 
No, it implicitly involves time
An horizontal slice would be:
A specific product in all countries for a particular year
Right?
 
9:58 AM
you could just look at all cell phones across time on all continents
(one sec)
 
All continents sorry :P
Is that the only possibility then in this diagram? A vertical and horizontal slice?
 
well you have three dimensions so you have three one dimensional slices and three 2 dimensional slices, and the data could be aggregated in other interesting ways too
Maybe it would be helpful if we looked at how the data would probably be stored in a star schema?
 
Yeah, that'd be cool.
 
Ok, so in this case we'd have a table for each dimension
one would store years
one would store product categories
one would store continents
With me so far?
 
Yeah
 
10:06 AM
Then we'd have a fact table which would store whatever information we needed on each of these dimensions.
Chances are the aggregation level would be a lot less
but we could have a row for cell phones sold to Asia in 2003 for example
 
So if we had 4 dimensions around a fact, the cube would be 4 dimensioal?
 
(more likely it might be per invoice, per product, or maybe aggregated per day)
yes
you can have as many dimensions as you like :-)
 
I think the suggested max is 20.. imagine trying to imagine 20 dims... good luck! Lol
 
so then what we can do is aggregate based on whatever dimensions we want.
with anything beyond three we aren't going to think about things as a physical cube
we'll probably just focus on the squares we are interested in for a given report.
 
Right, so this aggregation depends on the business process right?
 
10:10 AM
Yep. These will also almost never be involved in transactional workflows afaik. I have trouble imagining cases where they might.
 
Is it recommended for users to have all possible aggregations?
 
They will be used (almost?) exclusively in business analysis and management
I think that's a business call.
 
Makes sense
I'm feeling slightly better :D
 
Do you know of any online apps that shows a use of OLAP systems?
Maybe the cube itself
 
10:13 AM
hmmm
you mean visualizing OLAP cubes?
 
Or just how it is laid out.. I mean users don't use cubes ha
I was told not think of it in terms of actual cubes and tables..
 
right
which rdbms's are you planning on using?
 
But from the images I have seen.. it is laid out like a normal RDMS table
@ChrisTravers I'm not.. I'm just understanding the concepts in preparation for an exam.
Theory isn't usually enough for me :( I need images, applications, in depth explanations rofl. I'm that kind of person
 
That's good
 
It isn't :( It is very slow
 
10:16 AM
well, it means better understanding when you get there.
 
That's right.. but if I never use it I will forget and be back to square 1 :D
 
haha
or technology will move on and the understandng will be worthless!
(remembering the efforts I put into learning how to troubleshoot the LILO boot process on Linux)
 
And it was obsolete? Ow
 
(now it is, then it wasn't)
I am thinking of examples of star schemas, one sec
 
Is it possible to think of aggregate combinations as individual tables
Although the third dim doesn't fit
 
10:20 AM
This has an example that roughly matches your diagram
In general you want to set things up so you can aggregate on whatever criteria you want
if you want to find the number of cell phones sold in Europe in 2002, that can be done
The example there has a fourth dimension btw
(corporate organization)
 
I noticed.. ow :(
In that example the measure is sales of a product by date time right?
 
right
 
But sales can't be the aggregate
 
the goal is to ensure you can aggregate sales to look at it.
 
Mmn.. so where would I store the aggregate?
Oh.. so that (the aggregation) forms the cube (or whatever)?
 
10:25 AM
Ok so the fact table could hold one of two things
It will probably hold an aggregate of some kind
but the level of that aggregation depends on business needs
Chances are this will not be one row per unit shipped
 
I guess the question I am asking is.. can the fact table hold multiple aggregates?
 
it could be one row per invoice
Ah....
In general, I would recommend keeping any aggregates in the sales table very closely related
for example I could see having units sold and total value in a fact table together
since these are functionally dependent on the same things basically.
 
Yeah, that makes sense
 
(or more properly total value is functionally depend on number of units shipped and other data)
 
Right
 
10:28 AM
but you are pretty heavily constrained by your dimensions
 
Mmn.. but if the measure is.. something that isn't closely related.. I'm not good at business, can you think of something? :P
Maybe average?
 
if you tried to put rent expenses, then it wouldn't match your product dimension
 
Ah there we go
 
I was trying.... every example I could think of ran up against non-applicable dimensions
maybe....
 
So in that case we have a separate star schema ?
 
10:29 AM
or a separate fact table referencing only some of the dimension tables
for example we could have a fixed expense table or a fixed asset depreciation table.....
 
A star schema can have multiple facts then?
 
and it could reference everything but product
Each fact table is at the center of it's star.
 
Right..
I realized something.. a lot of DBAs have good knowledge of business concepts.. is it necessary (not that I want to be a DBA :P)
Stuff like sales, depreciation over my head
 
I write accounting software :-P
That's my primary thing
 
Lol.. I see.
 
10:32 AM
The DBA side is tied in because we put a lot of weight on the db.
I think one of the things you will find is that working with databases exposes you to business problems in a way that, say, network engineering doesn't
Business being business decision making and management problems
networking may expose you to operational problems in the same way however O.o
 
Dang, any tips on avoiding such stuff?
I hate business.
 
well....
 
I like software development. But business makes me sleep :(
 
the best way to learn it is to be self-employed :-)
 
A future goal is to learn to "accept" it.
 
10:34 AM
(and then you get to learn enough to run your own business)
Most of it is nothing magic, but accounting is pretty technical. You can almost approach it like programming.
Most of it is just knowing when to ask questions
and in terms of avoiding it, a good place to start would be working for a web business where you are involved more in service delivery and less on the accounting side.
anyway I have to take off for a bit. If you need anything leave a message here and I will get to it when I get back
 
Sorry, was talking to a classmate
@ChrisTravers Thanks dude! Very clear now :D Appreciate it. Take care :)
 
10:54 AM
26
A: How often do CAPTCHAs appear?

Jeff AtwoodTo trigger CAPTCHA while editing, asking, or answering: The minimum window is 5 seconds The maximum window is 40 minutes There is a maximum of: one edit every 30 seconds one answer every 60 seconds one question every 60 seconds To trigger CAPTCHA while doing performing other actions, ther...

> There is a maximum of:
> one answer every 60 seconds
O'rly?
2
 
11:33 AM
0
Q: Oracle transactions deadlock

user1545453How to rollback all active transactions on Oracle DB? I execute query and see 4 transcations in ACTIVE status. SELECT * FROM V$TRANSACTION

 
JNK
12:16 PM
@孔夫子 A lot of those Jeff answers I think are probably outdated at this point
he's been gone over a year I believe
 
 
1 hour later…
1:38 PM
Can't tell whether this is /s
+1 nice answer. oh by the way, thanks for the good exampleJW 웃 5 mins ago
 
@FreshPrinceOfSO strange I wonder what that is for
 
Please; stop adding unnecessary tags to questions. This just makes it look like you're doing it to get the badges. Myself and others will continue to remove them and roll them back if need be. Reference. — FreshPrinceOfSO 18 hours ago
 
JNK
2:02 PM
LOL
Take a look at his account page now
> Yes, I am always wrong. And you are always right. I am nothing but such a piece of nonsense tag editor.
3
 
oh wow
 
I guess our work is done
 
JNK
he also commented on jurgen
he is taking it personally apparently
 
@JNK link?
 
JNK
2:04 PM
which, unfortunately, is not surprising for someone who edits tags to get badges
 
+1 good answer, by the way I also knew the person you are talking here. — JW 웃 33 mins ago
 
JNK
+1 good answer, by the way I also knew the person you are talking here. — JW 웃 33 mins ago
 
It is funny that he commented on really old answers
I wonder where my comment is.
 
JNK
if he feels that strongly about it he should post in meta
 
2:05 PM
If you think XML queries are complicated stuff you should really have a look at the query plan for this query to get some perspective.
 
@MikaelEriksson That's pretty sexy
 
JNK
@MikaelEriksson nice
 
I'm suprised SQL Server got it all togheter
 
@bluefeet you're next.
 
@FreshPrinceOfSO I am sure. If he has issues then he should respond on Meta, that is what it is for
 
JNK
2:07 PM
I might be too since I posted a meta Q about it
 
ugh XML queries
 
Well @AaronBertrand edited all of the questions that the .
 
@JW웃 Do you disagree with the discussion? I don't see any comments from you. — FreshPrinceOfSO 6 secs ago
 
JNK
he's removing the bad tags himself it seems
 
@bluefeet ALL?
 
2:09 PM
@FreshPrinceOfSO all
 
@bluefeet okay that's hilarious
 
JNK
damn that should take a while
 
Here it is folks
 
JNK
there were like 900 of them
 
no, I agree and i have nothing to say. — JW 웃 28 secs ago
 
JNK
2:10 PM
@bluefeet no he didn;t
there's 700+ left
 
I edited a few this morning as well
 
@JNK he removed the tags from the questions JW answered
 
JNK
looks like he did the last few months worth though
 
@bluefeet these are really the only ones that mattered
 
JNK
no they ALL matter
it's not about JW it's about the craptastic tag
 
2:11 PM
12 hours ago, by Aaron Bertrand
http://stackoverflow.com/search?q=user:491243+[mysql-error-1064]
@JNK I agree, I was just saying that he removed them all from JW
 
@JW웃 Can you help clarify who your profile comment is directed at? — FreshPrinceOfSO 1 min ago
@bluefeet Looks like there's still more clean up to do..
 
Those are tags that have been requested previously. I don't think they should be removed
 
@JNK JW edited his profile..
> and Yes, I am nothing but such a piece of nonsense tag editor.
 
2:19 PM
11
Q: Organize tags for SQL key words

Erwin BrandstetterThere are a number of tags for SQL key words. And I think that makes sense, mostly. However, many of those keywords are not very selective or descriptive by themselves as they are very common words. As SO grows, many of them are competing with other fields. I started editing a new tag wiki on sql...

 
JNK
@FreshPrinceOfSO I know I pointed it out
 
@JNK he modified it.. after i asked him to clarify who it was directed at
 
JNK
lol
how passive agressive, this isn't personal
 
@JNK This is super silly
@bluefeet I think this is where we need to start.. stackoverflow.com/users/491243/jw?tab=tags
Still doing it.. sigh
 
JNK
I think adding SQL is fine
 
2:24 PM
@FreshPrinceOfSO I have no problem with adding . I am not sure about the
 
@bluefeet If it's tagged , should it have both and ?
 
@FreshPrinceOfSO I don't have an issue with that
 
Sprocs that call sprocs that call sprocs sweet!
 
states the platform
If the syntax is sql server specific, then I would add
 
@bluefeet keep going.. i want to try to follow your guidelines
 
2:26 PM
@FreshPrinceOfSO This shouldn't be just my opinion, what does everyone think?
 
JNK
I think that's fine
I think sql is fine if it's a querying question not specific to a platform
 
@JNK I agree but the remaining ones are the few added by the questioners, one here and one there. The only user who has been consistently adding silly tags in all the questions he answers, is that guy. That is still obvious in the tag: stackoverflow.com/tags/query/topusers
 
JNK
yes he adds dumb tags
but the problem is with the tags
 
@JNK And don't forget that DBAs like to automate tasks.
@JNK I don't disagree.
 
@ypercube I think the should be next. Awful pointless tag
 
2:33 PM
@bluefeet yeah. It is also used in LINQ questions and probably others. It's worse than mysql-error-1064, I think.
Anyway, I have work to do. See you later.
 
@bluefeet @ypercube are we burniating this tag?
 
JNK
query is bad too
 
@FreshPrinceOfSO there are 28k+ questions tagged with it. That will be difficult.
 
@bluefeet let's start with JW's
 
@FreshPrinceOfSO I don't think we should start with his. I think we should start from the back and move forward.
 
2:36 PM
@bluefeet I like your thinking
 
I know he is a pain with the tagging but we shouldn't solely focus on him
 
And I'll stop saying bad things for JW. He is a devotee now.
+1 this is awesome. — JW 웃 8 mins ago
 
I think he might feel ganged up on
 
I think he's following us now. Forget his cult.
 
JNK
yeah he is following some of us at least
 
2:39 PM
0
Q: Installing Oracle 11g on ubuntu 12.10

spitfireddI am installing Oracle 11g so I can work through practice excerises. I am following the installation walkthrough at link. In the terminal when I run chown -R oracle:oinstall database I get Operation not permitted How to I fix this so I can install the database?

prolly needs to be mig8ed
 
does sql server 2012 have a greatest() function?
before I comment that they are wrong I wanted to double check
0
A: Is there a Max function in SQL Server that takes two values like Math.Max in .NET?

LE GALL BenoîtDon't know why everybody try to re-implement the wheel, cause there is alreday a function as answer here which give you : SELECT o.OrderId, GREATEST(o.NegotiatedPrice, o.SuggestedPrice) FROM Order o

I didn't find any info when I bingled it but I might have missed it
 
@bluefeet according to google, no. results come back with top.
Hey SQL Server DBA's...
How does the process work for mapping a network drive to a database?
 
2:57 PM
@FreshPrinceOfSO more importantly, why do you want to do this?
In 2008 / R2 it requires a trace flag to even allow it, and even then it's not supported IIRC.
 
@AaronBertrand need to open an MS Access db from a stored procedure
 
So why do you need to map a drive? You can't use the \\UNC\path\?
 
@AaronBertrand It doesn't have permissions to access that drive.
 

« first day (841 days earlier)      last day (4019 days later) »