« first day (2201 days earlier)      last day (2666 days later) » 

12:50 AM
god damn
try to answer a question on SQL Server. Find it it doesn't have string aggregation
when does it end?! =(
 
 
4 hours later…
4:21 AM
@bluefeet yoo my frienndddd
What does blankman do?
Is that internal?
I am 99% certain the user is hard core gaming the system and something is way way up.
There is no way you can be that stupid and have that much rep.
Everyone of his questions is crap
Most of them are some seemingly automatic generation of official docs that replace specific identifiers with metasyntatic variables. Which, if true, is awesome.
If he really exists. He's a "Question Squatter" and that's always awesome. That totally undermines so much of the incentive system..
 
 
4 hours later…
8:43 AM
@EvanCarroll The next version of SQL Server by the looks of it msdn.microsoft.com/en-us/library/mt790580.aspx
 
9:06 AM
@AaronBertrand agree, or should I write ugree
 
10:02 AM
Hola
 
10:44 AM
@EvanCarroll in that (deleted) answer was not clear that the code was meant to work in SQL Server (and version v.Next). That's one of the reasons you got downvotes (and the "upgrade to Postgres" did not help either ;) The link from JamesAnderson above has the syntax.
 
 
2 hours later…
1:08 PM
Oooh, I hadn't realized.
I have hammer, too now!
 
@ypercubeᵀᴹ congrats!
 
@ypercubeᵀᴹ Only one more tag to match your 4 sided hammer avatar
 
@TomV hah, yeah. Although that is just the unfolding of a 4d cube in 3d. The 6d unfolding into 5d has 10 sides
It was the question you found that duplicate from Denny that I realized it.
 
1:23 PM
@TomV - I hope you don't mind, I reformatted the table in this answer
 
@MaxVernon No, thanks, it's much better
 
<pre></pre> can be used to format without syntax highlighting
my pleasure by the way, it's a useful answer
 
@MaxVernon Learned something new, thanks
And the question is better formatted now too
I wasn't entirely happy with it, but was disturbed by a recruiter
I shouldn't pick up when the number is from southend-on-sea :)
 
@TomV lol, recruiters are such a pain
So UPS says my notebook is "out for delivery" as of 4:27 this morning. I wonder how many people answer the door at 4:30am.
 
@MaxVernon I have absolutely no idea how you could enjoy that job, they must get a kick out of being rejected
 
1:31 PM
@TomV no kidding. They probably like being strangled while having sex, too.
Is the server running on a mobile phone from 2002? All kidding aside, there are a lot of possible reasons. Have you checked anything? — Max Vernon 28 secs ago
 
1:49 PM
^^^ handy!
and most likely, in a lot of cases, doing that will resolve the issue.
 
2:00 PM
could this be a question for DBA?
0
Q: WHERE IN clause in SqlServer with Millions of parameters - C#

Irfan HabibQ: I need to pass Millions of parameters in WHERE IN clause, below is the scenario mentioned: I've two database servers SourceDB-server & DestinationDB-server, I need to Sync data from SourceDB TO DestinationDB, so I'm generating query dynamically through C# code, final query is mentioned below...

 
2:11 PM
@Lamak I'd say so.
I'd pass in a table-valued-parameter, and join to it, but yes, it's not a terrible question.
 
@MaxVernon anyway, it already has some answers....
 
that stink.
 
2:46 PM
SQL Server on Linux is going to be much fun
 
Oooh snow
 
we had some snow, too. Yesterday and today
Temperature not low enough for the snow to stay much though ;)
 
Looks like it's going to stay in my garden, but definitely not on the roads
 
We've had snow on the ground since early December.
 
Looking forward to some drifting in the parking lot :D
 
2:50 PM
The digital thermometer was offline this morning. Which means it was damn cold, like at least -28°C
@TomV nice!
 
@MaxVernon cool!
 
@ypercubeᵀᴹ lol, very!
 
@MaxVernon How do you dress against that type of cold, jeezus
-5°C is considered cold here :)
 
@TomV not many people wearing shorts, for certain. Thermal underwear helps. I have one of these I wear to work:
plus a facemask
this:
its surprising to me this has only a single up-vote on the question:
1
Q: installed SQL Server on RHEL7 and received - Sqlcmd: Error: Microsoft ODBC Driver 13 for SQL Server : TCP Provider: Error code 0x2AF9

Fyodor GlebovI installed SQL Server on RHEL7 VM msodbcsql-13.1.1.0-1.x86_64 mssql-server-14.0.100.187-1.x86_64 mssql-tools-14.0.2.0-1.x86_64 # ls -al /opt/ total 4 drwxr-xr-x. 5 root root 52 Jan 13 08:31 . drwxr-xr-x. 17 root root 4096 Jan 13 07:55 .. drwxr-xr-x. 3 root root 22 Jan 13 08:31 microsoft ...

seems a very well written question to me.
our weather today has an "Extreme cold warning": weather.gc.ca/city/pages/mb-38_metric_e.html
 
@MaxVernon unfortunately the client tools won't install from the RHEL rpm on my fedora
the server itself did, surprisingly
 
3:01 PM
@TomV coool! I can't wait to try RHEL on my laptop.
in a VM, of course!
I was also thinking about getting docker to run linux on it. Never tried docker before, so that should be interesting/frustrating, I'm sure.
yesterday morning was "interesting" in Winnipeg: t.co/uK4OiySPWw
 
Could somebody please add [vnext] as a tag and tag that question?
+3 upvotes
 
@hot2use I considered it, but I'm not sure how useful that is. Consider as an example. What is "next"? I understand vNext is how Microsoft is referring to their upcoming version, but once it is released, what will refer to?
@hot2use thanks!
 
beta makes more sense than vnext
however there is still a maintenance problem
 
0
Q: Add tag for vNext?

Max VernonMicrosoft is referring to their upcoming version of SQL Server as "vNext". Should we add vnext? I think the transient nature of the word "next" indicates this would be a bad candidate tag since questions tagged with it will eventually not actually refer to the new "vNext".

@AaronBertrand agreed
 
issues that occur in a product's beta period may still be valid when the product is no longer beta. Who gets to go remove all the tags?
 
3:10 PM
@AaronBertrand I nominate @PaulWhite
lolz
 
@MaxVernon True. But people might start looking for vNext information and until the version is out, what better way to channel the google searches? You could then have a synonym refer to [sql-server-2017nix] or whatever MS comes up with.
 
@hot2use good point - maybe add an answer to the meta question
 
@MaxVernon just happened
 
It would probably still be more readable than SAP table names — Tom V 7 secs ago
:)
 
3:27 PM
@TomV amen!
 
3:43 PM
Interestingly, this page seems to imply tempdb does not get "rebuilt" on vNext; it states SQL Server will not start if it cannot "restore" tempdb.
 
4:27 PM
@MaxVernon i think that's poor phrasing (which wouldnt be a first for microsoft documentation)
 
@swasheck that's what i initially thought too, but it explicitly tells you to copy the tempdb files from the old location to the new, when moving tempdb. Which you don't need to do on Windows.
 
2
Q: Add tag for vNext?

Max VernonMicrosoft is referring to their upcoming version of SQL Server as "vNext". Should we add vnext? I think the transient nature of the word "next" indicates this would be a bad candidate tag since questions tagged with it will eventually not actually refer to the new "vNext".

 
it might easily be someone who just doesn't know how it works.
Bad @MasterDatabase!
 
 
2 hours later…
6:13 PM
HALLO AND GOOD MORNING
 
6:33 PM
WHERE ARE YOU?
 
@MaxVernon are you any good with math? i know @ypercubeᵀᴹ is ... but he's not around
 
who is not around? @swasheck
 
@ypercubeᵀᴹ i just had to summon you
@ypercube i've already got a solution using CASE, but i was wondering if you knew of any other way to work around the possibility of dividing by ln(1)?
(t-sql ... 2016)
 
6:51 PM
@swasheck What's the problem you're looking to solve?
 
ln(1) is 0 isn't it?
how can you divide? can you show us the CASE?
 
any variable to the 0th power turns into 1
therefore any log of (1) is zero
 
i know that.
 
oh, sorry ... /facepalm
 
is it something like ln(a) / ln(b) and you want to avoid errors when b=1?
 
6:54 PM
solving for 45 = 1^(3n-1) ... ln(45) = ln(1)*(3n-1) ... ln(45)/ln(1) = 3n-1
@ypercubeᵀᴹ yeah
 
Can you not just put a where clause on the b?
 
ln(45)/ln(1) = TTTTTHHHHHHHPPPPPPTTTTTTT
@jcolebrand i certainly can. just didnt know if there was other wizardry to avoid it
 
which means there is no solution to that equation
 
I thought the solution was Inf
 
I don't think so. CASE, WHERE. I can't think of anything else.
 
6:56 PM
ok. thanks.
@jcolebrand would you represent that as NULL? we have legacy databases with autogrow as a percentage and i've been tasked with forecasting when they'd hit maxfilesize
 
Depends on the usecase
 
(yeah. updated)
 
if you're calculating a limit (growth) then I would represent it as Infinity
But if you're calculating it as a size, then I would use null
 
@jcolebrand there are several types of Infinity. Usually we don't need arithmetic with them. Even those that we do (define and do arithmetic with), 1^SomeInfinity = 1
 
What you really need is three growth params, right? So it's sequential over time
 
6:58 PM
well i'm necessarily calculating at for which n (number of growths) we'd hit or exceed maxfilesize
 
(there maybe weird definitions that give different results but it wouldn't be nice theories)
 
In the case presented I would think that the intepretation of 1/0 is towards limits, so Inf
 
Guys, plus 7 is way too much for that agent vs ssms answer
 
I think there's a flaw in that equation
ln(45) = ln(1) * (3n -1)
45 = 1^(3n-1) is invalid
1^x = 1, always
 
@jcolebrand yes, for limits, the answer might be infinity.
 
7:00 PM
So the equation is naturally flawed, and can never be defined
Where did the 1^ come from?
or am I misreading an L as a 1?
 
@jcolebrand I suppose it's the growth factor. If it's 1, there is no growth, so no solutions to the equation.
 
Anyone futzed about with JSON in SQL Server 2016?
 
Which is a short circuit evaluation to begin with
When the case is 1 (gf) then you can't solve the problem
 
MAXFILESIZE >= (InitialFileSize)^((1*GrowthRate)*n - 1)
(sorry. had a work email)
wait ...
it's the InitialFileSize that's 1 and blowing me up
 
the edited equation doesn't make sense
 
7:04 PM
I think that this is better ...
MAXFILESIZE >=  (InitialFileSize * (GrowthRate ^ (n - 1)))
 
MAXFILESIZE >= InitialFileSize * ( Growthrate^(n-1 ))
yeah, we got thr same
 
aye
Try that @swasheck
 
Where Growthrate has to be >1 for the equation to have solution
 
ok
no. that's not correct.
 
What's not correct?
 
7:08 PM
that doesnt give the correct answer.
 
(MAXFILESIZE/InitialFileSize) >= GrowthRate ^ (n - 1) :: GrowthRate > 1
if GrowthRate = 1, the answer is always true
 
wait. hang on
 
Isnt' GrowthRate either a MB size or a percentage?
You shouldn't ever be doubling in size, so it should be something like 1.1 (for 10% growth)
Wait, I'm skipping a step
If it's 10% growth rate this becomes a recursive iteration
If it's megabytes then the equation above is correct, and you should fatally fail on values that are 1 so you can go do maintenance and fix them to be something valid
 
@jcolebrand i figured it out. i told you that GrowthRate is actual = 1+AutoGrowthRate and then didnt actually edit the equation ... edited and now it's correct. i think that's what i have originally, but didnt reverse-engineer the equation correctly here
 
No, you have it right. Fot 10%, the growthrate is 1.1
 
7:11 PM
Right, but then it's recursive, because you would be adding in the other amount, no?
Well, wait, thinking about the math
doing the power is the recursion
 
@jcolebrand which is why it's geometric and not arithmetic
 
@jcolebrand yes
 
because 1.1 * 3 = 1.1 * 1.1 * 1.1 = 1.21 * 1.1 = 1.331 or something like that
 
the equation we had is for % growth.
 
Right
 
7:12 PM
so long story short ... it's not solvable for cases when InitialFileSize = 1
 
And so in either case, 1 is an invalid case, because it will never grow
 
@swasheck it is
 
if i'm solving for n
 
@swasheck no, it is solvable for that case
(MAXFILESIZE/InitialFileSize) >= (1 + AutoGrowthRate) ^ (n - 1) :: AutoGrowthRate > 0
 
It is not solvable when the GrowthRate is 1 (if the AutoGrowthRate is 0)
 
7:13 PM
If the AutoGrowthRate is 0 then the answer is always true
 
ok. that's right.
thanks. derp
 
Because it will never grow
 
'@swasheck and what is the n? the number of growths?
 
future iterations to calculate against
 
@ypercubeᵀᴹ yes
 
7:14 PM
"if we grow n times, will we exceed the size"
 
then it should n, not n-1
or maybe no. Depends what you want to find.
 
right. because InitialSize = n0?
 
No ... that's not what you're solving for
^^^^^^^^
 
The maximum growths that is ok or when exactly problem wlil hapen
 
when the problem will happen
 
7:15 PM
Then you should be checking
(MAXFILESIZE/InitialFileSize) >= (1 + AutoGrowthRate) ^ (n - 1) :: AutoGrowthRate > 0
(MAXFILESIZE/InitialFileSize) < (1 + AutoGrowthRate) ^ (n) :: AutoGrowthRate > 0
So you need both scenarios
that's the trigger point, after n - 1 you're okay, at n you aren't
 
Log(MAXFILESIZE/InitialFileSize) / Log (1 + AutoGrowthRate)  will give you say 15.37. You can increase 15 times but not 16
 
@jcolebrand your second scenario is for non-percentage growths, which is necessarily arithmetic. i only need to check the first scenario, right?
 
@ypercubeᵀᴹ Yah, that's a better way of putting it haha
That gives the exact answer without having to do the math I was thinking of
 
thanks all. this has been a fun exercise in going back 25 years.
 
bye. see you later
 
 
1 hour later…
@bluefeet I feel cheap.
 
8:48 PM
@hot2use solomon added a comment to the connect item connect.microsoft.com/SQLServer/feedback/details/1959379/…
 
9:42 PM
@TomV thanks. I'll have a look.
 
 
1 hour later…
10:54 PM
@bluefeet interesting question
 
11:39 PM
Here is one.. If I'm right in my answer -- I may not be. I think that VERBOSE is super crap for postgresql in this case. dba.stackexchange.com/a/161042/2639
Let us never say that I was mute on PostgreSQL's warts.
 

« first day (2201 days earlier)      last day (2666 days later) »