« first day (4452 days earlier)      last day (415 days later) » 

12:34 AM
0
A: Varbinary update attempt

CharliefaceIf the column is declared as varbinary(max) then you can use the little known .WRITE syntax to do this. I have used a Nums function to generate the correct amount of rows. There are a number of different versions of this function out there, see https://sqlperformance.com/2021/01/t-sql-queries/num...

Have you ever used .WRITE in anger? Send in geolocated selfie to...
I honestly can't think of a single good reason to ever be doing this, but the customer is always right.
 
@Charlieface ...the customer is always .WRITE 😎
3
 
 
2 hours later…
2:10 AM
i'd be a lot more tickled to discover this emoji exists if windows had like... flag emoji support emojipedia.org/dino-cat
 
 
4 hours later…
6:28 AM
Morning
 
 
2 hours later…
8:38 AM
@PaulWhite yes
morning
Wordle 633 4/6

⬛🟨⬛⬛⬛
⬛⬛⬛🟨⬛
🟨🟩🟨⬛🟩
🟩🟩🟩🟩🟩
 
9:36 AM
Wordle 633 4/6

⬛🟨⬛⬛⬛
⬛🟨⬛🟨🟨
⬛🟩🟩🟩🟩
🟩🟩🟩🟩🟩
 
 
2 hours later…
11:46 AM
Wordle 633 3/6*

🟩⬛⬛🟨⬛
🟩⬛⬛⬛🟨
🟩🟩🟩🟩🟩
 
12:35 PM
🟩🟩🟩🟩🟩
🟩🟩🟩🟩🟩
🟩🟩🟩🟩🟩
🟩🟩🟩🟩🟩
 
dork
 
@PaulWhite donk
 
@ErikDarling DOINK
 
dink
 
🤾‍♂️
 
12:46 PM
are you calling me the white artist formerly known as prince?
 
Well that works just great on Windows lol
@ErikDarling It's a basketball player on iOS
🧺🏀
I forget which character you used to use for "dunk"
 
oh
🤾‍♂️
 
That's the fella
 
sort of looks like sanskrit
 
fixed
"man playing handball"
 
12:50 PM
"before being beheaded for losing at handball"
 
It's a lot to pack into a few pixels, isn't it
I can sort of see how Chinese is such a compact language
Z is bound to have some sort of appropriate simplified/traditional meme for that
 
who, @Zikato?
i remember when @Zikato had time for blogs and memes
now @Zikato is lost in the @Forrest
 
Oh yeah, I forgot his full handle for moment, yes, @Zikato
Forrest must be a really tough boss
 
you know his bark is worse than actually no i'm not doing this
 
There's this sort of trend where people start to prefer paid work
I appreciated the bark pun btw
 
12:59 PM
🙏
 
I'm glad you decided to leaf it be
Oh no
 
here we go again
has anyone noticed a weird question
 
hah
Do you know, I almost considered installing my first ever AG today to look into that
And as if by magic, a Sean appears
 
did he ever return, no he never returned, and his fate is still unknown
 
So the regulars in here would be @ErikDarling, @HannahVernon, @PaulWhite, @Zikato, @J.D., @JoshDarnell, @SeanGallardy, @mustaccio, @Forrest, @Charlieface, @CadeRoux, @TimStone, @Lamak, @mustaccio and @billinkc according to the listing at the top right.
 
1:02 PM
@ErikDarling That user has never left a comment
Ever
 
you may not like it, but this is what peak user looks like
 
There should be some sort of award for that
Gosh yes, a Peak User badge
 
meta is your friend
 
So this chat has peaked?
 
years ago
 
1:03 PM
@SeanGallardy Does that pique your interest?
 
it's all decline now
 
@PaulWhite only if it's high school
 
@ErikDarling and then fall
 
So we should change the name of the room to The Peak™ – Consultancy ©®
 
Though we call it autumn
 
1:04 PM
It's time for spring cleaning, though
 
when rome falls, so falls the world
 
Time to buy a 🎻
That's a violin/fiddle
 
Viola!
 
a comment from @Charlieface has more upvotes than my great answer©
is it time to retire
 
@SeanGallardy Did Nero viola while Rome burned?
 
1:07 PM
@JohnK.N. I'm an irregular regular :-) and I only found this place recently
 
@ErikDarling He'll never be a peak user
 
@ErikDarling which one?
 
1
Q: Why my clustered index causing deadlock (Keylocks and Pagelocks)

Jishnu ChandranWe are experiencing deadlock 2-3 times a day in SQL Server. Upon analysis, we have found that the deadlock is caused by a combination of keylock and pagelock occurring simultaneously. We are actually using EF to fetch data and the query involves more than 6 foreign key relations. But on some days...

 
@PaulWhite I don't believe Nero was that with it
 
@SeanGallardy It's hard to say, and he's not here to defend himself. Another one we'll never hear back from
 
1:09 PM
@ErikDarling Answering with a guess about what the query plan might look like, and me commenting please supply the plan? Yeah I can see why you might need to retire...
 
Not sure he'd have much of a defence
 
it's not a guess
 
It's weird there's parallelism
 
You're probably right anyway, but Paul isn't going to forgive you.
 
Eh? What would I need to forgive Erik for?
 
1:10 PM
i pay my tithes
 
It's a guess in as much that we don't know which of the queries has the key lookup, or why there is parallelism (most likely those two are correlated).
 
@ErikDarling Most prompt invoice payer ever
 
@PaulWhite For guessing an answer? We can guess all day long but "I never guess. It is a shocking habit,—destructive to the logical faculty."
 
well what else would cause an object-level shared lock in a select query
 
Oh well 'guesses' based on experience and expertise are fine. Guessing at an answer in a comment is the thing
Unforgiveable
 
1:13 PM
the deadlock xml is a disaster, parallel queries log an entry per thread or something
 
@PaulWhite That's an educated guess, for which one needs an education. Explains why I never do it :-)
 
I have 3 votes left today
learn.microsoft.com/en-us/sql/relational-databases/xml/… Clearly those users are using the wrong SET options. Why is a different question, you would have to check the software they are using. There shouldn't be any good reason to use anything different than the docs anyway. — Charlieface 2 hours ago
 
@ErikDarling Could be a different statement in the same transaction, for a start. Could be an indexed view update, could be an ID column which isn't actually unique
 
is betting allowed here?
 
I think I'm going to flag that as Not A Comment
 
1:15 PM
@PaulWhite We've had this before: I'm happy to guess in a comment, because a guess is not an answer. I've no idea where else to put "Not An Answer"
And it doesn't get to the root cause "why?"
 
The site needs unofficial not answer, answers
 
Or maybe a "community answer in progress" where people can contribute. Oh wait...
 
@Charlieface Why do you persist with this. We've had this debate. Local policy is firmly in favour of not putting guesses at an answer in comments.
 
only the op can choose the answer from what is submitted
comments can't be marked as answers
 
14
Q: Is the comment-only answer acceptable?

Peter VandivierI've just read a post that's really mundane. It's one of those "just read the error message" type questions. There's an obvious, correct, and incredibly trivial answer I can respond with; but I'm not sure the original post is high-quality enough to deserve a full post in reply. So... Is the comme...

I really don't get it. If the answer turns out to be wrong, just delete it
If there's clarification needed, ask for it
At the very least an answer gets its own comment section rather than cluttering up the question box
@ErikDarling On a separate note, I always think SqlWorldwide is your sockpuppet
 
1:20 PM
i may need to start drinking earlier than usual today
 
Oh, there's the problem: I have no drink
 
no excuse for that
 
I'd make a guess if comments were downvote-able like answers, there would be a percentage of less comment-answers.
 
comments should cost reputation to leave
2
 
Don't give them ideas
 
1:24 PM
@ErikDarling mm even better hah
 
Could we rename comments to something like sub-answers?
 
@J.D. Yeah, that's been suggested many times on main meta. If you go that route, comments become ever more like answers. Just leave an answer and be done with it.
 
@JohnK.N. and what if a sub-answer needs clarification?...we should add a comments feature to it. 🙂
 
Threaded comments are also a popular request.
 
@J.D. Good idea...
 
1:25 PM
(not with me)
 
@PaulWhite I remember seeing them in the wild somewhere. Can't remember where though. Some old Q&A site....
 
@JohnK.N. Codidact has them, I think
 
@PaulWhite That's surprising to me. I'd figure the opposite for 2 reasons. 1. Comments feel like less of a commitment than an answer because they are more simplified and hidden from general view, and can't be downvoted. So it's ok (from the commenter's perspective) to be incomplete, partially inaccurate, or straight up incorrect - without risk of repercussions.
 
that would require a reworking of the comments table to have a parentcommentid and then aaron would have to write 17 posts about recursive ctes and adding columns with zero downtime
 
There's also a certain arrogance attached to answering in a question comment. Like your thoughts are so important they need to be un-downvotable and positioned above even an accepted answer.
 
1:27 PM
ROFL
 
2. People tend to clean up their own mess when they see it's gathered a general negative response, which downvotes represent. Also the fact it hurts some people's internet points too.
 
@ErikDarling You're talking about a company with questions and answers in the same table
 
@PaulWhite That's one of many ways of looking at it, if you're specifically looking for it
 
Well, everyone has bias, it's true
 
not moderators
 
1:32 PM
They sort of fudged the issue by only having the top x comments visible. I've said many times over the years, having an off-screen or side area for comment-chat on the question would be acceptable. But we don't have that. Or self-expiring comments.
 
impartial to the core
 
It's our motto for sure
 
> DECLARE @SQLDate varchar(8000)
 
@ErikDarling Is there any other way?
 
@ErikDarling It's wrong to punch people isn't it
 
1:38 PM
not that i'm aware of
 
Could've been worse, at least not VARCHAR(MAX)...
 
Sean's never had a date that long
DECLARE @SeanDate tinyint;
 
womp womp :(
 
@ErikDarling Sounds like the guy who was writing his own ORM and wanted to use VARCHAR for everything.
 
i wonder how he's doing
 
1:42 PM
Still being strung along
 
I couldn't CHAR less?
 
Oh, here we go with the var-car thing again
 
flagged as abuse
 
@PaulWhite I guess that requires commitment to the answer being right, because downvotes. If I have a guess at an answer, I'm going to tell OP "by the way maybe try this". Also requires commitment to writing up a full answer, rather than just a one-line half-baked sentence. Fact remains, there's no way currently to send the OP on the right track unless you do either a comment or an answer, and personally I find half-baked answers far more irritating than half-baked comments.
TBH I think the culture on Stack Overflow is a bit different also, where I'm more used to comments.
 
I don't find that persuasive. If comments could be downvoted, you'd have exactly the same concerns. If an answer gets heavily downvoted, sure, you might choose to delete it to rescue your score from those beastly -2s, but in my experience decent guesses don't tend to attract DVs much because they might help someone else in future (and answer DVs cost 1 rep).
If you don't want to write half-baked answers, great! Just don't do it in comments either.
And yes, we don't look to SO as the gold standard to emulate.
We don't apply the rules without consideration either. If someone has voted to close a question and left a helpful 'pointer' sort of comment too, well, fine.
 
1:58 PM
what level of inactivity does a stack site have to hit before it gets closed down
 
Oh really astonishingly low
It's only ever happened a few times
 
interesting
 
Like no non-spam activity of any kind, and no moderator team for six months or so
 
so i can't vote to close dba.se
 
One problem with the slack approach is that once a question disappears off the front page, people forget about it. They don't follow up on old comments, and the site just gets cluttered up with old, unanswered questions with a few comments that went nowhere.
 
2:01 PM
which comes back to the comment expiration date feature
 
Would've been worth a shot, but the idea didn't get a lot of general support
People were worried about broken comment chains or something as I recall
But mainly the wider network is quite resistant to change or experimentation
And the company has been too, unless it clearly benefits SO (the site) or Teams revenue or whatever
Fair enough, to some extent, seeing as they are a for-profit enterprise
 
think about all the money they could save on disk space by deleting comments!
3
 
For anyone wanting to follow the meta rabbit hole: meta.stackexchange.com/questions/linked/204402?sort=active
@J.D. How is any of that a good thing for the site?
 
did you get a drink yet?
 
No, hence a bit grumpy
 
2:14 PM
sort yourself
 
I also need to clean the kitchen
But someone is wrong on the internet
 
always
wait are you talking about me
 
No, you're always correct, and a prompt payer
 
tiger woods y'all
 
Ok, I'm going to clean the kitchen
 
2:18 PM
expecto sandwichum
 
2:33 PM
this guy really had me going
-1
Q: How to ORDER BY after two tables have been joind using INNER JOIN ? ? Please answer without aliasses

michal roeslerHello Database Admins. I have the database and within it one table is called dbo.tbl_Cities and it looks like this: CREATE TABLE dbo.tbl_Cities (City_ID INT IDENTITY(1, 1) NOT NULL, City_Name VARCHAR(70) NOT NULL, City_State_Region VARCHAR(100), ID_Country TINYINT NOT NULL CON...

 
Does anyone know how to get sp_executesql to return non-zero without it throwing an exception?
Docs say
> Return Code Values
>
> 0 (success) or non-zero (failure)
 
does using an output parameter count?
 
No an actual return value. As in EXEC @i = sp_executesql ...
You can't use RETURN 42; you get A RETURN statement with a return value cannot be used in this context.
 
oh, no
never really tried though
 
2:55 PM
@ErikDarling Apparently any non-batch aborting error will return the error number dbfiddle.uk/DAHHju82
Bit difficult to see in db<>fiddle because it only shows the error message if there is one.
 
> CREATE TABLE Regyptian (Strut int NOT NULL CHECK (Strut > 0))
wat
 
Don't ask, I was copying straight out of Erland's article
He's got some whacky names in there and I thought you were the funny one.
This one is batch aborting and has no result dbfiddle.uk/u3Iy0Y_i
 
i thought i was funny until i met you
 
3:58 PM
@PaulWhite That's my point. That's the current state of comments. If they were downvote-able, then they'd be more of a commitment, similar to answers, to follow the purpose of them.
 
Well ok, but comments are different from answers for good reasons
 
For sure, I'm not debating how comments should be used. Just the mere fact that a way for the community to police them the same way as answers, would seem beneficial to me.
 
Here's some background information: blog.stackoverflow.com/2009/04/… ... blog.stackoverflow.com/2009/04/comments-top-n-shown - a lot has changed since '09, but that should give you a rough idea of the reasoning behind this system. Read those, think about them a bit, look over a bunch of different posts with comments on The Site Of Your Choice, and then work through a few different scenarios in your head... Then come back with an idea for how down-votes might actually affect the behavior of the system and examples of specific problems solved. — Shog9 Mar 13, 2013 at 16:29
Plenty of rabbit hole available for the interested observer, or those who feel they have a new and important point to make on meta
 
4:22 PM
On the specific point, regular users can already police comments through flagging. Just three users flagging an ordinary disposable comment is enough for it to be deleted without a moderator becoming involved, for example.
 
user image
4
For @SeanGallardy who doesn't use Twitter
 
Nice, we can expect weekly memes again?
 
no
@Forrest keeps me too busy
 
if you think about it all of chat is a comment section
 
4:27 PM
without downvotes
 
4:48 PM
@PaulWhite Fair, I never noticed the Flag icon before. But also no tangible repercussion to the user, since all it does is just undo what they did.
 
Don't forget the changing buttons in chat
 
@J.D. Having a comment deleted is not a tangible result? I mean, having your question or answer deleted also "just undoes what you did" but...
 
5:09 PM
have you ever saw someone use centiseconds? Like instead of 10ms to say 1 cs
 
Yep, just now
About 6000 cs ago
Otherwise, no
 
hmm, too bad
 
In the US, they probably prefer fahrenseconds
3
 
how many eagles per football field is that?
 
Four, rounded up
 
5:27 PM
@PaulWhite Sure there's a result, but nothing that would make most people think twice about posting crappy comments again in the future. Same thing for answers, get rid of downvoting and only allow flagging and removal without explanation. I would bet you'd see the same people willing to continue to post crappy answers, even more so than in the current system.
 
Maybe. Luckily, we're still small enough that we can recognise crappy posters and commenters by name or avatar most of the time
And eventually, someone would raise a custom mod flag, or we would see one of the comment flags, see the pattern, and take an action
 
No doubt.
 
People with low quality contributions don't tend to last more than a few hundred fahrenseconds in general
Couple of thousand, tops
 
That's a lot of Eagles..
 
5:42 PM
No dogfood for Victor tonight
 
5:54 PM
@SeanGallardy I was rereading your post and found a typo
https://www.seangallardy.com/cluster-resource-dll-connectivity-in-availability-groups/

"Yes, but the best thing is to not shoot yourself in the foot for not reason. Leave the default ports."

For **not** reason instead of **no** reason
 
6:06 PM
@PaulWhite don't talk about me like i'm not here
 
@Zikato He'll be devastated. Nothing Sean hates more than spelling errors
 
6:19 PM
@Zikato I think Oracle trace reports timing in centiseconds
 
@mustaccio You've convinced me
I'll never mention centiseconds ever again
 
Wordle 634 5/6*

🟩⬛🟩⬛⬛
🟩🟨🟩⬛⬛
🟩⬛🟩🟩🟩
🟩⬛🟩🟩🟩
🟩🟩🟩🟩🟩
Really nailing those 5s at the moment
 
Spoken like a true oracle
 
Anyone have any ideas why the wizziwig editor is bugging out on the T-SQL code I put in on this one? - dba.stackexchange.com/a/324761/150011
 
⬛️⬛️🟩⬛️⬛️
⬛️⬛️🟩⬛️⬛️
⬛️⬛️🟩⬛️⬛️
🟩🟩🟩🟩🟩
🟩🟩🟩🟩🟩
🟩🟩🟩🟩🟩
 
6:29 PM
@ErikDarling 👏
 
@Zikato Good catch! Thank you :)
 
Seems to be due to the ordered list that comes before it. Had to put a silly HR between the list and the code block.
 
6:48 PM
Human Resource?
 
HR = horizontal rule = <hr />
I think that's a long-standing bug, but I'm too lazy to look for it 😁
I think it works if you use code fences rather than indentation to format the code.
 
does anyone have a good template for using a cursor with error handling to keep going when errors are hit?
like lets say i have a list of views to create, and some views have dependencies on other views
i want to try to create all of them, but if i hit a dependency error just keep going to the next one and then retry others later after the other views are successfully created
 
7:41 PM
@JoshDarnell Thanks, I was wondering if it was a bug. Are code fences the backticks symbol before and after each line?
 
60 centiseconds from disaster
 
sir please use valid freedomheit units
v. unwelcoming to the best country in the world
 
8:11 PM
@J.D. Backticks on a separate line before and after each code block, yeah.
512
Q: Implement ```-style (fenced) Markdown code blocks

Mathias BynensCurrently, Stack Exchange’s Markdown parser only allows four-space indents to represent code blocks: // some code // another line of code GitHub Flavored Markdown and other Markdown implementations allow for an alternative syntax, that doesn’t require indenting each code line: ``` // ...

And since I was on Meta already...
168
Q: Code block is not properly formatted when placed immediately after a list item

Stephan202Consider the following piece of Markdown code: This is some regular text. >>> def factorial(n): ... return 1 if n < 2 else n * factorial(n - 1) ... * This is a list item. >>> def factorial(n): ... return 1 if n < 2 else n * factorial(n - 1) ... Notice that ...

 
8:31 PM
@ErikDarling the views are created with dynamic sql?
 
@Zikato yes
 
9:08 PM
@ErikDarling I don't know how to run only one batch in fiddle, but try this

https://dbfiddle.uk/qE64JcwE
 
that doesn't create the view with the dependency after it creates the view it needs though
i may have to do it in two steps
 
it does if you keep running it in loop
ah, I see what you mean
 
9:25 PM
@ErikDarling dbfiddle.uk/8SteNESp
It had to be a temp table though, dynamic doesn't work with table variables it seems
 
oh that's a great idea
thanks
 
I'll take that win
 
Shocking: A for-profit company to stop offering terabytes of disk space for free
Next in the news: the earth is not as flat as previously thought
 
@JoshDarnell That worked, cheers! 🤙
 
10:20 PM
No problem!
 
10:32 PM
@mustaccio i keep hearing that disk is cheap
 
In other news, anyone looking for a new job? My company's hiring for fully remote positions now. lol
 
10:52 PM
@ErikDarling that's got to be more about making people pay, no?
 
Who's ready for some CUs?
Oh, another open source thing that is doing something people don't like and will break all their stuff with 30 days notice. I was told be reputable sources open source was the best in everything for everything of everything 🤷‍♂️
 
@J.D. if they are data related let me know and i can ask around
@SeanGallardy for what?
 
11:28 PM
@ErikDarling oh word, I appreciate that! We're mostly looking for full stack or more application layer geared, but if you know anyone that's more proficient on the database side and has some level of application layer experience, that could work.
 
> keep hearing that disk is cheap
probably from the same people who want others to pay for it
they may not realise you need some expensive boxes where to put those cheap disks, in some expensive buildings, with expensive security, fire protection, and air conditioning
 
11:46 PM
@HannahVernon the best way to get people to pay is to send them invoices
@mustaccio you mean the cloud?!
 

« first day (4452 days earlier)      last day (415 days later) »