« first day (384 days earlier)      last day (4492 days later) » 

 
3 hours later…
2:12 PM
1
A: Should our FAQ draw a line for "DBA" SQL questions and "SO" SQL questions? Is there one?

Jeff AtwoodI think this is a great idea -- if you haven't already, I suggest a diamond mod editing this addendum into the DBA FAQ. http://dba.stackexchange.com/faq I do see but this is not the right place to ask questions about... client-side programming (ask on Stack Overflow instead) basic...

 
 
1 hour later…
3:19 PM
whoa whoa whoa, what?
I'm curious to see what the outcome of this is
 
> Hope he doesn't have something dangerous on him like a copy of the constitution
Heh
 
Can't see this through the firewall. What's he just done?
 
Ron Paul
My son Rand is currently being detained by the TSA at the Nashville Airport. I'll share more details as the situation unfolds.
As far as I've ever been able to tell, that's about the most clean young man in Congress
 
I'm sure he's fine. Your chance of being detained by the TSA is inversely proportional to your chance of actually being dangerous best I can tell
 
3:40 PM
My son @SenRandPaul being detained by TSA for refusing full body pat-down after anomaly in body scanner in Nashville. More details coming.
@BenBrocka except when you refuse patdown
 
People who refuse patdowns are at least 99.9% people who have read the 4th amendment rather than terrorists. To date I'm not aware of their "enhanced patdowns" actually finding anything
 
3:57 PM
@ConcernedOfTunbridgeWells hello :-)
 
We have a proposition for you lot here
922
Q: FAQ for Stack Exchange sites

Justin StandardCommunity FAQ For sites in the Stack Exchange 2.0 network To see a list of commonly used words and phrases, see the glossary. Content Asking questions How do I write a good title? How to ask a smart question? How can I get answers fast? Is there a limit on how many questions I can ask? ...

If you're not familiar with this Q, it might be a good idea for you to peruse it
 
After the BI merge we thought it might be good to have our own version here, linked to by the faq, with more detailed information about what we want migrated, what "DBA" really means etc etc :-)
 
Give us a chance to break things down neatly, with a simple link from the FAQ, allow us to keep the FAQ as clean as it is now
Also, everybody should watch this ted.com/talks/…
 
@JackDouglas Hi
 
4:20 PM
@ConcernedOfTunbridgeWells thought you might be interested in this conversation :-)
 
4:34 PM
I love you guys.
 
For what?
 
This problem is why I prefer Oracle mandating all SQL statements terminate with a semicolon.
@jcolebrand Just in general. Thought I'd say that in case I died suddenly.
 
@NickChammas lmao, ok
I love you too man
 
@NickChammas do you also find us amusing ;-)
@NickChammas Oracle doesn't mandate that btw :-)
 
4:44 PM
@JackDouglas I find you add value to my day in many forms, among them certainly is amusement.
 
but I know what you mean
 
@JackDouglas No?
 
eg in SQL Developer you need a ; between statements but if you only have one statement it doesn't need it
it's a client tool thing rather than a database thing anyway...
 
@JackDouglas yeah, certainly; in that user's case, the two statements would have been interpreted as one and caused a syntax error
if that was on oracle
instead, the poor guy emptied his table
 
technically doesn't the SQL spec require a semicolon between queries (even contiguous ones) and the engine parsers are just smart enough to generally know where semicolons belong?
or at least, the grammar
 
4:47 PM
@jcolebrand yah, that's how SQL Server rolls
but MS is warning everybody to start using semicolons because they will become mandatory
 
@NickChammas nasty. like when I pressed F5 in SQL Developer again the other day after messing around with SSMS.
not blaming SSMS mind
PICNIC
 
> ;
> Transact-SQL statement terminator.Although the semicolon is not required for most > statements in this version of SQL Server, it will be required in a future version.
 
@JackDouglas problem in compiler, not in chair, I agree
 
that's what that means?
lol
 
@jcolebrand you are too kind :-)
 
4:50 PM
@NickChammas it means "Problem In Chair, Not In Computer" ;-)
I prefer PEBKAC
 
@NickChammas no
 
Problem Exists Between Keyboard And Chair
 
aaaaaaaaaaaahhhhhhhh
 
oops too slow
 
PEBKAC is the one I've heard
@MarkStoreySmith on page 3 now... wtf
 
4:55 PM
@NickChammas Scary stuff!
 
@MarkStoreySmith nobody uses ALTER DATABASE BEGIN BACKUP though
that's ancient history
the whole article is overblown hype IMO
 
@JackDouglas You still went and checked your SCNs though didn't you? :)
 
you're kidding - this is a DW, we get about one SCN per minute :-)
5170830546
 
230 trillion odd to go for you then!
 
5:10 PM
ha ha :-)
 
From what I understand from the article, the problem is really confined to very large deployments with linked systems right? As the SCN is cascaded between linked databases?
 
omg
2
Q: Is my bank storing my password in plain text?

fredleyMy bank (and every bank I've come across) only ever asks for individual characters from my password when logging in. Is my bank storing my password in plain text?

 
@MarkStoreySmith They are saying if you use an obsolete method of backing up your database deliberately in an RAC cluster there is an outside chance of an intractable problem. To me it looks like "if someone evil has access to one of your databases he could sabotage them" - which is kind of a truism.
 
@NickChammas The two UK banks I use, 2 cc providers and investment platform I use all do this
 
ALTER procedure [dbo].[Dash_ControllableExpenses]

as

select
	'2009-11-30' EndingDate
	,200000.00 MonthlyExpense
	,1800000.00 YTDExpense
That is the completeness of this query
WTF?
 
5:15 PM
@jcolebrand no bugs in that baby
 
@NickChammas no documentation either
 
@MarkStoreySmith isn't this a bad, bad sign for their industry?
 
I'm not even sure where the hell to look to see what it's for
 
@NickChammas The answer from jimbob covers why its used, to prevent replay attacks
 
@MarkStoreySmith this is good question :)
 
5:17 PM
It's just picking on the poor sad lonely Oracle people ;-)
 
@MarkStoreySmith don't see why it needs to be reversible encryption
 
@JackDouglas Do you want to rephrase the FAQ in terms of a series of questions (per a traditional FAQ) with links out to deeper discussions of the individual questiions?
 
you have hashes for all (n choose 3) possible combinations; why should they be reversible?
 
And maybe add summary answers to the questions, e.g.
Are questions about reporting tools in scope for this site: Yes (with a link out to a meta question that discusses it in some depth).
 
@NickChammas Not an expert on such matters but I guess that by storing hashes for a combination of 3 character combinations, it becomes viable to brute force the hashes
 
5:23 PM
@MarkStoreySmith I'd like to know how that is. Just asked the dr
@JackDouglas 3,711,639,348
 
@ConcernedOfTunbridgeWells no nothing that drastic - just add a brief subsection at the end of the top section of the faq with a link to a meta Q like the one on mSO Cole linked
@NickChammas thats potentially a lot of combinations ;-)
 
6:00 PM
@JackDouglas - answer updated to include max SCN
 
@NickChammas but I held off on my downvote anyway ;-)
 
@JackDouglas lol
i hit submit and then I was like, wait... I didn't answer my own question completely
if anybody knows how to pull the SCN for <9i versions of Oracle, please add your answer
 
Anyone know if SS or other RDBMS can do select distinct pk_leading_column without a full scan?
 
you're talking about a multi-part PK
?
 
6:18 PM
@NickChammas yes that's what I mean
but I'm not really talking about keys - just whether the optimizer can use leading columns of an index (including non-unique index) to do a distinct assuming all are not null.
I am not good at communicating my thoughts :-)
related to comments on an answer of mine here
 
7:01 PM
altering events, even alter event evt_name enable; causes it to become disabled on slaves. kinda WTF IMO
 
JNK
7:22 PM
@JackDouglas it's going to be a clustered index scan no matter how you dice it I think. It needs to read the whole clustered index but I think it can use the stored metadata to get those distinct values more quickly
@JackDouglas - just confirmed. I made two 3-column tables, all fields are varchar(10), with identical 1m rows each. I indexed table 1 on A,B and table 2 on B,a. SELECT DISTINCT ColA runs in around half the time, with half the CPU, on the one with the index on that field FIRST. Opposite holds true as well for SELECT DISTINCT ColB
 
@JNK cool
 
 
2 hours later…
9:22 PM
is this guy right?
27
A: Is there any reason to worry about the column order in a table?

AndomarColumn order had a big performance impact on some of the databases I've tuned, spanning Sql Server, Oracle, and MySQL. This post has good rules of thumb: Primary key columns first Foreign key columns next. Frequently searched columns next Frequently updated columns later Nullable columns last...

 
the edit is right
> EDIT: Sql Server 2005 stores fixed-length fields at the start of the row. And each row has a reference to the start of a varchar. This completely negates the effect I've listed above. So for recent databases, column order no longer has any impact.
modern engines know how to fix screw ups, same as compilers
 
oh whoops, yup
that's what I though
 
9:49 PM
for those of you using C#
0
Q: How do I keep my stored procedure inputs from being silently truncated?

Nick ChammasWe use the standard System.Data classes, DbConnection and DbCommand, to connect to SQL Server from C#, and we have many stored procedures that take VARCHAR or NVARCHAR parameters as input. We found that neither SQL Server nor our C# application throws any kind of error or warning when a string lo...

 
 
2 hours later…
11:20 PM
morning chaps
we don't have any Sybase experts kicking around, do we?
 
11:35 PM
there's this guy I know who lurks on here
his name is Simon Righarts
he knows some Sybase
 
Haha. Was looking for an expert, not the village idiot. :D
2
... oh wait.
 

« first day (384 days earlier)      last day (4492 days later) »