« first day (1444 days earlier)      last day (3416 days later) » 

1:29 AM
@swasheck HAHAHAHAHAHAHA so true man
 
2:06 AM
 
@swasheck I had the reverse problem while in the US - trying to find something that approximates to a flat white - the ubiquitous coffee here.
 
indeed. i'm still trying to grasp what exactly it is which is owed to my lack of tactile experience with it
 
@billinkc A bigint isn't big enough (maximum value 18446744073709551615). A decimal(20) would be fine though.
 
@PaulWhite we decided it would be easiest to just not worry about fiddling with the intermediate types and go straight to XML
 
Oh ok, still reading forward in time :)
 
2:17 AM
does anyone else find it incredibly arrogant that mongo calls its server process mongod?
@bluefeet whats up
 
@swasheck modding, what's up with you?
 
i'm trying to work up the motivation to finish a blog post. just ... cant ... do ... it
 
sounds fun. I've been working on one for about a year but life has gotten in the way
 
things are more difficult for me than for someone like @PaulWhite who just mind-melds with the optimizer and asks it what it wants to teach humanity about itself
 
Oh yes because writing is that easy. I never go months without being able to write. Er.
 
2:31 AM
heh. i know - i'm just exploring my inferiority complex
 
One thing I am still learning is not to set the bar too high for oneself.
Or to be dissuaded by the apparent high quality work of others.
Fully accepting that the next post probably isn't going to change the world is harder than it sounds.
Which is not the same thing as saying the world needs more Pinal Daves.
 
of course. i'm less dissuaded by the high work of others, and moreso by the fact that i tend to go into uncharted territory (for myself) so everything is essentially an "alpha" version for me and who knows what's going to be wrong with it :)
 
Just try to avoid writing complete crap.
:-D
 
@PaulWhite I feel like I've written maybe two posts like that in my life, and still felt they weren't taken as inspirational as I predicted
 
@AaronBertrand I can relate.
 
2:38 AM
decidedly on-topic:
 
But whatever. If one person appreciates the post and learns from it, two thumbs up. Same as giving a presentation - I don't expect to enlighten everyone in the room, but I'd only be disappointed if I enlightened nobody. I haven't always been this way - see my post about feedback from my first SQL Bits, I took a couple of scores pretty hard.
Now I'm much more careful about writing a good title and abstract. If people show up but don't get what they expected, I'm far less to blame than I used to be.
 
that's a good point.
 
@swasheck I read that, pretty much, and am none the wiser.
I never read evaluations any more, even though I know I should. I gave a talk once where I got 100+ gushing evals and one bad one. It took me months to get over it.
 
@PaulWhite a very bad blog post about a very bad blog post :)
 
@swasheck Sure is!
 
2:44 AM
so meta
 
2:56 AM
@PaulWhite Do you define an unsigned long as 2^32-1? Otherwise, how can it not fit in bigint 2^63-1
 
Modern definition of long, I guess. A long long in old speak.
 
Ah, I was going with this definition arduino.cc/en/Reference/UnsignedLong
 
Understood.
 
But I was also assuming it was free form XML and they were just interpreting strings as whatever data type
 
One never knows on SO.
 
2:58 AM
Troof
Presented for your entertainment, possibly
Nevermind, that was not as good as their usual offerings
So many people think they can drive rally. So many examples of them not
and with that, off to play some mindless video games instead of rehearsing a presentation
 
 
2 hours later…
5:01 AM
 
 
2 hours later…
7:01 AM
Hm, does this person really know much about SSIS or XML?
0
A: XML file datatypes and sql equivalent for file import to sql

Paul WhiteThe SQL Server type decimal(20, 0) is large enough for the maximum xml unsigned long value, 18446744073709551615. DECLARE @x xml = N'<e>18446744073709551615</e>'; SELECT @x.value('/e[1] cast as xs:unsignedLong?', 'decimal(20,0)');

 
doubtful
 
7:45 AM
@AndriyM He knows XML well enough. Just don't like it much.
 
Ha!
2
A: Help with SQL Query to XML output

Mikael ErikssonIf you do 'NewID' as [field/@id] you will get a field element with an id attribute. On the next line you add B.ID as [field] ot get the value of ID as the node value to the field node you created on the line before. <field id="NewID">1</field> After that you want a new field node and to create...

 
8:12 AM
Can't help it. I find stuff like this immensely interesting.
XML variables with the "exact" same content has different performance depending on the way XML was loaded to the variable.
 
@PaulWhite we really should grab a beer sometime. 'cept I'm heading up north for two weeks on Sunday
 
8:28 AM
Hi from Zurich
3
 
8:39 AM
@SimonRigharts Yes we really should. Ping me when you get back. Where you off to?
 
@PaulWhite Mahia - north end of Hawkes Bay, then Napier. Two weeks of lounging around, food and drink, sunshine and beaches. It's gunna be tough. :)
 
@SimonRigharts My sister lives in Napier, so I know how tough it can be up there.
You might actually get to see some summer!
@ypercube Hi! Making a deposit? ;)
 
8:56 AM
@PaulWhite Summer? What is that?
 
9:37 AM
@MikaelEriksson Seems to be the same root cause as an issue I've written about before
Converting strings to XML over and over and over again is expensive.
 
10:26 AM
@PaulWhite I don't think so. It has to do with how the data is structured inside the XML variable.
A repro to play with
declare @N1 varchar(50)
declare @N2 int
declare @X xml

-- 100 row, 10 item per row
set @X =
(
select top(100)
       V.name,
       V.number,
       (
       select top(10) V2.name,
                      V2.number
       from master..spt_values as V2
       for xml path('item'), type
       )
from master..spt_values as V
for xml path('row'), root('root'), type
)

declare @D1 datetime = getdate()

select @N1 = R.X.value('(name/text())[1]', 'varchar(50)'),
       @N2 = R.X.value('(number/text())[1]', 'int'),
The two levels of row/item with the cross apply is necessary to get any difference.
So something to do with how nested item nodes is represented within a row node in the internal format of the XML datatype.
 
@MikaelEriksson Ooo yes. So converting a generated xml type to nvarchar(max) and back is a good general workaround? Seems so.
 
Or leave out the ,type in the for xml query.
 
Well yes.
 
Then you get a warning in the query plan about an implicit conversion to XML.
 
Indeed.
 
10:39 AM
Same thing. The XML loads from a text representation.
 
@MikaelEriksson But you could suppress by introducing explicit conversion, right?
Or would such a warning be a non-issue?
 
@AndriyM I would not worry about the warning and if you do cast you would still end up with a fast version of the variable when omitting type. The cast to XML does not muck things up.
 
I always like to be explicit about type conversions.
 
Morning
 
10:52 AM
Hi
 
11:06 AM
@PaulWhite That may be because you like your query plans to be perfect, as if they were your dear children – because you deal with QPs often. Any warning in them would probably seem a blemish to you: "Ugh! Where did you get that? Here, use this CAST immediately!"
2
 
@AndriyM Ha. Yeah. Actually it's more just plain bitter experience. I have wasted so much time and effort over the years diagnosing invisible implicit conversions and related problems, I try very hard to avoid the things in the first place.
6
 
I find that with DW architecture. In many cases I'd rather have a design that's not necessarily optimal for performance, but is easy to test and built to avoid certain failure modes by design.
You can really see the value of design in defect rates. When I were a wee lad straight out of university I was on a VB6/SQL Server project. I built the SP layer for it so it was easy to unit test, and that whole subsystem had just two defects in about 13k lines of T-SQL stored procedure code.
 
Exactly. The right sort of effort up front pays huge dividends later on.
 
By comparison, the VB6 layer that I did had about 35k lines or so and the defect rate in that part was an order of magnitude higher (we had no technology to do auto unit tests on the GUI).
I had a similar experience at my gig last year. I developed an ETL process and a change capture framework (CDC on the source system wasn't available). About 3/4 of the issues in that turned out to be deployment as opposed to issues in the code. That process was designed to be easily tested and validated, with change capture done using a generic module that could be tested separately.
Performance wasn't great - it took about 3/4 of an hour to run as it had to process a bunch of complex queries over the entire source system, then compare the results to identify changes.
However, it was fast enough, pretty robust and easy to test and reconcile.
And it didn't take all that long to build - maybe a couple of months in total.
 
 
3 hours later…
2:00 PM
morning crickets
 
morning :-)
 
@bluefeet how dare you dare me dare you call me a cricket
 
I think @Kermit has lost it
 
I had chocolate for breakfast #imanadult
 
My fave is cake for breakfast or ice cream
I've had both as an adult
why are you not wearing a hat @Kermit?
 
2:12 PM
@ypercube Hi from Lake Zurich
@bluefeet i only participate in questions
 
lame
 
lol
 
it's a vertica life
@Lamak did you ever hear back about the CM role?
 
nope
 
@Lamak bastards
 
2:18 PM
not even a "thanks for participating"
 
@Lamak not even a hat?
 
yeah, damn bastards
@Kermit I should demand one
 
@Lamak i'll support your campaign
 
2:55 PM
thanks man
 
We are completely becoming SO
it says incorrect syntax on .key parts of SELECT 1 FROM dbo.target AS t WHERE t.key = s.key would t.IDENTITYCOL = old.IDENTITYCOL accomplish the same thing? — Dust 22 mins ago
 
@AaronBertrand It's because I keep telling people to come to DBA. :P
 
@AaronBertrand woot woot
 
3:10 PM
I made an analogy that even Jeremiah appreciated
 
you should use a porsche though
 
@Lamak i wonder where brent will park it
 
he doesn't need to, he just leaves it there and buy another one
 
@PaulWhite Yeah. Instead of buying hats. So many internet points ;)
 
3:29 PM
How do you know it don't work? — billinkc 7 mins ago
 
@Kermit I should've said "and buys another one", right?
 
WTF
so what i should use now !!! i need to have a column inside the database to check for concurrency conflicts ,, so should i use timestamp or rowversion ? and if i use timestamp ,, what about the MSDN link which mentioned that i should not use it, and if i want to use rowversion i can not !!! so what is the solution ? — john G 3 mins ago
 
@AaronBertrand he asked the same question on SO
 
@AaronBertrand some people just want to be dicks, regardless of the mental gymnastics required to achieve the goal
 
0
Q: Cannot add “rowversion” data type in sql server 2008 r2 management studio, or even to modify existing column data type

john GI want to create a column with Rowversion data type, but inside my SQL Server 2008 R2 Management Studio, I cannot find such a data type: Also when I tried altering existing column named timestamp which is of type timestamp using this statement: ALTER TABLE [SkillManagement].[dbo].[Customer] A...

I mean, a related question
 
3:33 PM
@AaronBertrand want me to migrate the SO version to me merged?
 
@bluefeet yes please
 
but it isn't exactly the same question. From the SO one he got that timestamp and rowversion are the same
 
@Lamak the information he got from comments / answers does not change the question.
 
yeah, that's right
 
3:40 PM
And merging doesn't remove any of that. It just eliminates the redundancy of the question itself.
More fun... where do people even get the idea that they should use something like deleted.%%physloc%%???
0
Q: How to filter out deleted records in a trigger on a table with composite primary key (MSSQL)

Louis SomersI am trying to optimize a trigger on a table in a database from a 3rd party. I am not allowed to change or add any columns to the schema. The table has a composite primary key, one column is an integer, the other is nvarchar. How can I filter out the deleted rows without using a temp table? I ...

 
I just read the merged question, seems good
 
currently we use sql server 2008 r2, but we might decide to use sql server 2012 or maybe newer sql version in the future so i want to make sure i do not face problems ... why you mentioned it does not matter if i use timestamp or rowverion ? — john G 28 secs ago
FUCK. Read.
^^^^ actually don't do the first. Let's keep the gene pool strong.
 
Is the underlying issue they have a column with a name that is a reserved word?
 
yes i did .. you mean even if i use rowversion , sql serve will map it as timestamp... so shouldi keep using timestamp ? — john G 14 secs ago
@billinkc possibly, but that is only causing an error when he tries to alter the column. The thrust of his question (and 99% of his comments) just seem to be seeking advice for things he hasn't absorbed yet, like why he should use timestamp when the documentation tells him not to.
 
1
Q: Can I sacrifice consistency to allow concurrent 'insert or update' and read access to a single table with SQL Server?

Collin DauphineeI have a table that contains two columns: a resource key, and (very roughly) when it was last accessed. I have a number of servers that are periodically dumping data about resource accesses to the table. They should either update the access time for a resource key if it already exists, or insert...

isn't MERGE the answer here? ^^^^
 
3:46 PM
@JackDouglas I got so excited, I thought that was going to be a meta question about whether we can sacrifice SO users
 
so. much. teeth. pulling
-2
Q: How to combine multiple SQL Queries into one output

user2240189I am having troubles combining multiple queries into one output. Here is my code: SELECT SUM(CASE WHEN answer = 1 THEN 1 ELSE 0 END) AS [Male] FROM answers where questId = 2 SELECT SUM(CASE WHEN answer =2 THEN 1 ELSE 0 END) AS [Alcoholic] FROM answers where questId = 6 SELECT SUM(CASE WHEN a...

 
@billinkc you should post one on M.SO
 
i think that it's funny that you folks still try to answer questions
 
Tempting. But the skin has just started to grow back
 
Yes, let me repeat some things I've already stated multiple times. SQL Server will map rowversion -> timestamp. They are the same thing. If you are using the GUI, you don't have a choice. If you are writing CREATE TABLE scripts, you should use rowversion IMHO, but it's not going to make any difference on current versions of SQL Server, because - again - rowversion just gets mapped to timestamp under the covers anyway. Did I cover everything? Again? — Aaron Bertrand ♦ 12 secs ago
 
3:49 PM
@JackDouglas MERGE has issues
 
@AaronBertrand some people need to take a reading comprehension test before being allowed to post questions
 
oh, yuckity yuck
 
@Jack But it is not single, atomic operation unless you force it to be so using elevated isolation. Ahd there are worse problems, too. Please read this post. I actually persuade all of our customers to avoid MERGE and have even phased it out in a few places where it has been used because of these issues. — Aaron Bertrand 1 min ago
 
thanks @Aaron
 
3:51 PM
MERGE makes implicit promises it can't keep unless you know how to force its hand.
And there are a bunch of lingering bugs that I certainly wouldn't want to be responsible to regression test against.
The old school method may be a few more characters but I can rely on it.
 
why would you do this?
0
Q: Changing a column type and keeping the data

Shems EddineI have a sql server 2005 database where one of the columns is an int. The database is full of data so I can't lose any of the data. I need to change the the of the column from an int to a varchar(50) but this column is used in a lot of stored procedures where they are assigned to variables as i...

why do I feel that @AaronBertrand's head is going to explode today?
 
@billinkc do you actually use that color combination in SSMS?
 
I use something similar but a bit less feminine
 
4:07 PM
how can someone have that much trouble understanding the issue?
but if it is required to have rowversion then i should stop using the designer? second question is this a limitation insdie sql server 2008 R2 management studio , beucase inside my sql server 2012 express there is a datatype named rowversion ... — john G 2 mins ago
 
Who knows. It's like talking to a f*ing wall.
@johnG NO IT DOES NOT MAKE ANY DIFFERENT. — Aaron Bertrand ♦ 14 secs ago
 
VtC --- tip of the iceberg. we can't solve your tangential issues of reading comprehension and basic thought
 
@johnG how many times do I need to say that rowversion gets mapped to timestamp? Are we going to do this all day? At what point will you believe me? — Aaron Bertrand ♦ 9 secs ago
 
@AaronBertrand disengage. he's not worth the heart pills
 
I wonder if they're trolling
 
4:10 PM
it's actually brent trolling aaron with a fake account
4
 
That would be a Christmas gift that would make the Grinch's heart explode
 
@swasheck I wish he'd do it on SO!
shall we migrate it there :)
 
Wish I could lock an answer from subsequent comments
 
please
 
@AaronBertrand you mean without locking voting/editing?
 
4:15 PM
Right. Or even block a certain person from commenting.
 
That moment when you are writing up an answer and realize you just wrote enough to making it into a good blog post...
 
I consider SO my blogging engine
2
 
do we want this on dba?
1
Q: SQL Server 2008 R2 - Change Tracking, Database Snapshots, invalid rows

KrystoffI have a system in place where a SQL Server 2008 R2 database is being mirrored to a second server. The database on the first server has change tracking turned on and on the second server we create a database snapshot of the mirror database in order to pull data for an ETL system with the change ...

 
4:31 PM
@billinkc I use SO/SE because I can't be arsed maintaing a blog
 
You're like the brother I never wanted
 
Indeed, the cool, hip brother who was drinking flat whites before they went mainstream.
 
I drink jaded black to match your soul
 
Are you implying that my decade or two working in I.T. might have made me slightly cynical?
Not quite glass half full?
Perhaps even 'not a team player?'
 
I do find that I am far more cynical about employment and employers after having been burnt.
 
4:45 PM
@billinkc Yes. I used to be ever so bright eyed and bushy tailed. However, I've been burned a number of times too.
Today's Economist has an ad for network switches on page 7. Of course the switch is called 'CloudEngine' as opposed to something meaningful like 'network switch'.
 
Huh, underwriting data for life insurance policies. They store this data ... poorly
Floats everywhere. Column name ties back to a MetaData table. ALQ130 is clearly the number of alcoholic drinks per day in the past 12 months
I'm sure the app they use keeps all this nice and tidy with some ORM
We must have Wilt Chamberlain's policy here. Number of sexual partners, 99999
 
Today SAP is timing out. Thankfully we have a SAP consultant here to insist the problem must be index fragmentation on tables with fewer than 1 million rows.
 
Just imagine if it was 1.01 million rows
 
@billinkc Most policy admin systems have pretty crap data quality.
 
Besides the increase in licensing fees
 
5:04 PM
Anyone going to visit the Guantanamo Bay gift shop?
 
@Kermit Is that where the embassy is going to be?
 
@James makes the most sense
 
stupid excel
 
it really is the worst.
 
CANT EXECUTE STORED PROCEDURES WITH PARAMETER PROMPTS!!!!!!!!!
 
5:17 PM
It sounds like you're trying to make it performant. Perhaps you meant SELECT * FROM ALL_TABLES
 
@swasheck why are you executing sps from excel?
powersomething?
 
PowerSomething, the hit comedy about a team of independent professionals trying to balance the ability to deliver for their customers without sacrificing their servers. Coming to you, every day of your life
 
you should make it happen
 
@Lamak trying to empower the end-user
 
ah, see?, you should give them nothing, but take from them EVERYTHING
 
5:29 PM
@Lamak At last, you are ready for America
 
finally
 
Remember, if the baby is not strong enough to hold onto its lollypop, you're allowed to take it
 
but the baby might then shoot me
 
The risk is worth the reward
 
that might be true
 
5:43 PM
@Lamak you could always steal the babies gun first.
 
yeah, I have the whole thing planned out
first, I cover the sun
 
Erg I'm trying to come up with a better way to automate our data lineage and it's proving tricky.
 
6:32 PM
How do you guys track that sort of thing.
 
What is data lineage?
 
6:56 PM
HOW IS BABBY FORMED?
 
JNK
How do I shot web
 
I accidentally all the data
 
F this noise I'm going to eat some lasagna.
 
screw you guys, i'm goin' home
 
MS had announced a project out of MS Research that was going to be the cool data lineage tool. It has since been expunged from the web. Seems it's a hard problem to solve well
Whoa, I just made the big leagues
> I wish to discuss a very confidential project with you
 
7:22 PM
@billinkc welcome to the recruiter spam show
 
Was that intended to be a reference to "Independence Day?"
 
eh?
 
Well done Comrade
We will show those Capitalist dogs who is boss
 
Can't wait to get some Cuban cigars
 
7:26 PM
@Kermit you and everyone else
 
@bluefeet should i take a number?
 
faux news (money ed.) is calling it an attack against capitalism
 
@Kermit yes probably. But you can exclude me from that list
 
@bluefeet i'll +1 you
 
I will take up smoking just to spite the two of you
 
7:31 PM
how noble
 
I do look rather regal in this velour smoking jacket
@swasheck Data's scientist character: welcome to the freak show
Also, I can't find diddly on youtube anymore
 
oh yeahh
 
Implied foreign keys are the best
 
7:53 PM
@billinkc That sounds like something an evil DB clippy would say.
@Zane You can use some repository tools. EA can do it at an individual field level but the UI to enter it is really clicky and tedious. However, there is a fairly good COM API that would let you load mappings directly into the repository from a source like a spreadsheet.
 
EA?
 
Powerdesigner comes with a data lineage/ETL specification metamodel but it's an order of magnitude more expensive than EA
@billinkc Enterprise architect.
If you're feeling really keen - as in keen enough to go rolling your own ActiveX controls - you can build extensions to EA that hook into its GUI.
However, for the amount of sodding around needed to do that you might be better off just getting PowerDesigner.
X88 Pandora also has a good data lineage facility. You can make up any sort of hacky, convoluted process within the tool and it has a facility that will convert it to a minimal 'normal form'. However, Pandora is about £15k per seat.
 
15 pounds per seat, that's not baaa. Oh. I should really get my eyes examined
 
Most people just use Excel, although it's really clumsy for anything more complex than straight 1:1 mapping.
Could you do anything clever with BIML?
 
They've got new things in the tool for modeling and such. I can't remember whether there was a lineage feature
To be honest, when people go all nuts over data lineage, I usually glaze over as it's never seemed a compelling feature. Perhaps I misunderstand the usage
 
8:03 PM
Depends on what they mean.
If you're doing a mapping reference document so people can see the data flows then you need some way of documenting it. This sort of thing is quite a big deal in financial services so you can prove you know how your management information is derived.
 
If it's not tribal knowledge, it's no good
(for job security)
 
The other meaning used with this term is essentially tagging each row in your data with a data source and source key and making sure that the grain of the data is preserved so you can trace individual DB transactions back to source.
Essentially column-based data lineage vs. row-based data lineage.
Maybe @Zane could comment about the specifics of what he's trying to achieve.
 
8:18 PM
Must be some fabulous lasagne
 
@billinkc Indeed. It beats out data lineage.
 
I just followed a link to learn what exactly a "tea length dress" was. Don't know that I have a clue after reading the "article" but I regret the ads that are sure to follow
 
@billinkc Is this a British term?
 
I wouldn't think so. It was used in a mailing from my boy's school about the charity auction being HS Prom themed which implies a certain vintage
 
8:34 PM
@billinkc alternative to a kilt
 
Nope, a kilt is shorter than a tea length dress, which should hit about mid calf
whereas the kilt should just barely brush the ground if you were to be kneeling
 
i'm just telling you that you can change up your skirts if you want to try something different
 
I would be interested in a ... pacific island garb. I think it was Samaon. Similar concept to a kilt
 
9:18 PM
If anyone wants to GIMMIE DA CODEZ, I'm open to it
Nevermind, I'm biting the bullet and using charindex
 
9:43 PM
soooo ... updates included in CU1 are also included in CU4 (SQL 2014), right?
 
yes, cumulative updates
 
this is horrible: mis.fortunecook.ie
 
@AaronBertrand brandon just made the same point to me. i just wanted to make sure i wasnt making too large of an assumption
 
@billinkc Is this what you mean? sqlfiddle.com/#!3/2e065/8
 
Yup, looking at your to see if I like it better
 
9:53 PM
@billinkc It's called a Lava Lava
 
@billinkc I am sorry your data model is so terrible
 
@AaronBertrand So say we all
My original and lazy solution was to add the key into the value. It was deleted and recreated in the source system but we can't delete from the DW because ... because. But since that "number" needs to be unique, we have to get the old one out of the way
 
weird
 
Thus, we'd have had 210012.999.100 00000000-0000-0000-0000-000000000010 etc in there but noo, that's not user friendly
 
A lava-lava is an article of daily clothing traditionally worn by Polynesians and other Oceanic peoples. It consists of a single rectangular cloth worn as a skirt. The term lava-lava is both singular and plural in the Samoan language. == Customary use == Today the fashion remains common in Samoa, American Samoa, Tonga and parts of Melanesia and Micronesia. It is worn by men and women in uses from school uniforms to business attire with a suit jacket and tie. Many people of Oceanic ethnicity wear the lava-lava as an expression of cultural identity and for comfort within expatriate communit...
 
9:55 PM
At which point, I think I should go to the bank so I can wipe the tears away with their money
@ConcernedOfTunbridgeWells Appreciated. Wasn't sure if you were trolling and suggesting I should wrap myself in lava. Although there is a video of a person stepping on lava and not being Smeagol'ed so I guess for very quick exposure, it's a risk I'm not willing to take
 
OMGWTF?!
> FIX: Parallel deadlock or self-deadlock occurs when you run a query that results in parallelism in SQL Server
> This issue occurs when parallelism is allowed at server or query level.

(Process ID 85) was deadlocked on thread | communication buffer resources with another process and has been chosen as the deadlock victim. Rerun the transaction.
 
@billinkc I can probably find you a shop somewhere in South Auckland that will sell you one for about 20 kiwi pesos.
 
@swasheck I wasn't trying to be chicken little here but sqlblog.com/blogs/aaron_bertrand/archive/2009/03/21/…
 
@ConcernedOfTunbridgeWells I'll worry about it whenever I visit
But thank you
 
10:01 PM
@AaronBertrand reading the CU hotfix descriptions is horrifying
 
Closer to $NZ30 than $20 but close enough.
 
The lava-lava got me to en.wikipedia.org/wiki/Pe%27a which has a photo of a heavily tattoo'ed man's backside in it.
 
@swasheck you know how an ostrich might deal with that problem...
 
That'd look pretty freaking awesome but one tattoo is plenty. Plus, I'd be the big white poseur doing something like that. Also, not in shape
 
10:04 PM
@AaronBertrand fortunately, the actual descriptions can talk that ostrich off the cliff - except the parallelism description
 
Also why are you installing CU4? You know CU5 has an additional 56 fixes, right?
 
i'm not. i'm just looking through all of them.
i hadnt gotten to cu5 yet :)
 
oh ok
 
yay
 
@billinkc Plus, no-one here really wants to see your bottom - with or without tattoos.
The coating started coming off the touchpad on my computer the other day. I found a howto about replacing it with stick-on vinyl - and bugger me if it doesn't work quite well.
 
10:47 PM
I have no idea what these people are doing, but they managed to bork my permissions in the dev cube environment daily. And completely lock up the service such that it takes a reset to right it
 
My guess is they are doing a project deploy
 
11:05 PM
ugh. i need to turn off automatic disk defrag on my sql servers, right?
 
at least for drives that have database on them.
 
prozactly
 

« first day (1444 days earlier)      last day (3416 days later) »