« first day (2277 days earlier)      last day (2582 days later) » 
00:00 - 18:0018:00 - 00:00

12:52 AM
"New Zealand man challenges environment minister to a fist fight over water pollution policy"
 
 
2 hours later…
2:43 AM
I'm finally deleting some horrible SQL code I wrote a year ago. You are the only people who might be able to appreciate the work that went into this.
I present to you: DST-aware datetime conversion in T-SQL!
3
Of course it only converts between Pacific, UTC, and British time
(Why did I do this? Because there was a potentially super-urgent business requirement that we didn't know if we'd have to start meeting with NO notice at all. Before any C# code was written, I needed a query that could mostly do the job, and, I did it.)
And when I say horrible, it doesn't mean it's badly written or doesn't work or isn't a bit clever. It means it should never have been written in the first place. Or should never have to have been written in the first place.
But when you're 9 years late complying with some requirement or other, scrambling does tend to occur.
 
3:10 AM
Anyone old know where I can find sql-86?
 
@ErikE Interesting. Not on SQL Server 2016?
We have some TZ related issues too. Still not fully solved
 
3:41 AM
@JoeObbish Just 2012.
 
 
3 hours later…
6:15 AM
good morning
 
Good morning
 
Good morning !!!
@ypercubeᵀᴹ eerrrhhh ... good question -_-' (sweating my arse off) music to soothe potentially embarassing question -> youtube.com/watch?v=r6HjMgKqGRk
 
6:40 AM
Good Morning.
@EvanCarroll No sorry. But what is your definition of old?
 
1 message moved to Trash
 
7:00 AM
@EvanCarroll want to pay? iso.org/standard/16661.html
 
 
2 hours later…
9:20 AM
7 messages moved to Trash
 
9:35 AM
1 message moved to Trash
 
Lots of trashing going on these days
 
9:59 AM
@JackDouglas that's pretty neat
 
10:22 AM
@ypercubeᵀᴹ does it require coalesce just in case sum() = null?
 
10:50 AM
@McNets not if you have else 0
 
@ypercubeᵀᴹ, ok, now I understand jpmc comment
 
11:18 AM
@Marian glad you like it — let me know if you have any suggestions
 
@McNets you can answer the question: dba.stackexchange.com/questions/168595/…
 
12:19 PM
@ypercubeᵀᴹ you gave me a link time ago about it, I saved it but now I cannot find it. I've searched into the chat, but I can't get any result.
 
12:44 PM
0
A: Where can I find the first standardization of SQL, SQL-86?

McNetsDigging into google I've found some references about it from ANSI/ISO official sites. According to ISO: ISO 9075:1987 Information processing systems -- Database language -- SQL But it seems this documentation has become obsolete and has been retired. According to ANSI: ANSI IN...

 
@ypercubeᵀᴹ I've checked the gentleman's answer and tbh, it is not usable
as an answer for the question I've raised yesterday's eve
should I closed it and raise it again with DBA SE?
 
@McNets your link is broken (the last one) and are you sure it's legit copyright-wise?
 
@TomV sorry, my fault, edited. I'm not sure about copyright. Please have a look. You cannot download it.
 
@ypercubeᵀᴹ we've found the issue actually :)
 
1:10 PM
Does Postgres allow: 'row_number() over (order by (select null))'?
 
@McNets why would you do such a thing?
also, I don't think it works
 
@dezso to get row_number() according inserted order
 
@McNets there is no such thing
2
it's up to the executor to return tuples in any order it likes
 
46
A: What does "ORDER BY (SELECT NULL)" mean?

usrROW_NUMBER requires an ORDER BY clause syntactically. You cannot use it without one. SELECT NULL is a hack to shut up the error while not enforcing any particular order. In this case we don't need to enforce any order, so the fastest option is to use SELECT NULL. The optimizer sees through this ...

 
it turns out you can do SELECT row_number() OVER (ORDER BY (SELECT NULL::integer)) FROM (VALUES (1)) t;
but then, I believe, it's the same as SELECT row_number() OVER () FROM (VALUES (1)) t;
in any case, there is no guarantee this will return stuff in any order
3
 
1:17 PM
yes, I think it only works for MS-SQL dbfiddle.uk/…
 
1:31 PM
It's nice when people don't really try. Russkiy : First Question / Second Question
3
 
2:01 PM
@McNets it doesn't work for any RDBMS
19
A: In SQL Server, can I guarantee an order without an explicit ORDER BY clause when an index seek is forced on a table with only a clustered index?

Paul White Am I right that it will guarantee the order in this case without an order by clause? No. A Flow Distinct that preserves order (allowing ORDER BY without a sort) is not implemented in SQL Server today. It is possible to do, but then many things are possible if we are allowed to change t...

@McNets no need to DROP TABLE btw
@McNets here's proof: they aren't returned in inserted order here: dbfiddle.uk/…
 
2:33 PM
In SSMS, is there any way to alias 'p' as '['?
This would double my speed writing queries.
 
[robably but then [osting answers becomes a real [ain in the [osterior
7
 
someone looked into this more a little while ago
the number of values you're inserting can change things
also you inserted into a heap
definitely shouldn't expect order then
but if you insert enough rows into a table with a clustered index then it'll be (usually) inserted in order
 
@JoeObbish unless parallel query?
 
2:53 PM
@McNets I had liinks only for SQL-92 and 2008+. Not for 86
 
@JackDouglas SQL Server doesn't support parallel inserts into a table with a clustered index
 
@JoeObbish I meant parallel query
 
I always see the data be sorted then inserted in a serial zone
I guess I'm not quite sure what you mean
the pattern I see is that you do your work in parallel to find the data
then sort in parallel
then switch to serial zone
then insert in clustered key order
 
I don't know how SQL Server works, but isn't the clustered index a btree structure?
with all the leaf-block splitting etc associated with btrees and inserts
and if you are scanning the whole table, isn't it scanned in on-disk order rather than traversing the index tree?
 
depends
there's something called an allocation order scan which will scan the data on disk without caring about the clustered index order
 
2:58 PM
like an FFS on Oracle?
 
there are certain conditions that need to be met for SQL Server to consider that to be safe
 
:36388958 :p
 
it can also scan the index by traversing the index free
hm... I thought a FFS went in order?
 
which will be horrible for performance on a big table
 
at least I've written code that assumed that and it appeared to work... maybe it didn't...
are you thinking about fragmentation?
 
3:00 PM
In any case, I think we all agree there is no easy way (or no way at all) to get rows according to "insertion order". Right?
 
maybe with a trigger?
maybe someone could use %%physloc%% or whatever it is
I would say neither of those are "easy"
 
@ypercubeᵀᴹ "Insertion order" is a meaningless concept. Why was when you inserted the data important? You could argue that it means you're recording an event that occurred at that time, but that's what a datetime column is for. What if you discover an error in a feed and values are delivered late? Now "insertion order" does NOT represent the time of the event.
 
it can matter in oracle for IOTs
definitely not a clear concept though
 
@ypercubeᵀᴹ The best you can do is use a sequence, but don't believe it means the order the events occurred in.
 
@ErikE I agree 100% on that (it is meaningless concept)
 
3:03 PM
A database exists precisely, and ONLY, for answering questions about the past.
Even if you're planning future events, it's a record of your past plans.
Aside from the data's meaning, what useful question actually needs to ask "in what order did the data enter the database"?
This can only be a proxy for the REAL question which is, "what order of events does the data represent"?
 
JNK
@ypercubeᵀᴹ If you have a CreateDate column with default of SYSDATETIME() you can order by it on the select
 
@JNK yeah, no argue there. If you do need that, you have a created column, with values provided either by the db or the application.
 
Using the "physical" representation of the data store—a necessary but unimportant implementation detail for answering questions about the past—is to mistakenly conflate the activities of the storage mechanism with the meaning of the data.
 
Sorry for going all soap-boxy
 
JNK
3:06 PM
@ypercubeᵀᴹ but you're basically saying without planning for it, based on existing metadata
 
"One does single block IO and reads the index keys in sort order. The other does multi block IO and reads the index keys as it hits them, NOT in sort order."
 
oops
thanks for the correction
 
@JoeObbish I'm struggling to demo this on dbfiddle — possibly it only behaves like that when real multiblock io is occurring, ie the blocks aren't cached
 
@JoeObbish in Postgres it's the ctid - but when you update something, it changes, moving the affected row out of the order
 
3:23 PM
@dezso Oracle has row migration too but it doesn't happen on every update
 
@JackDouglas related, I think it was either MartinSmith or PaulWhite or both that mentioned a feature in SQL Server Enterprise where a query can use cached data from other session queries.
 
@PaulWhite hey. I was wondering if you know of a good place to read (and maybe test) about the new features of vNext?
 
@Lamak Not aside from the obvious: reading the documentation; downloading it or using e.g. Azure.
 
3:28 PM
@PaulWhite @JackDouglas I was referring to this one: Advanced Scanning
 
@PaulWhite yup, was what I was doing right now. But I thought that I had nothing to lose by asking (aside of your time, of course)
 
There's a technet virtual lab for installing vNext. Unfortunately the installation instructions are for Ubuntu and the provided vm is Red Hat. As I am a complete Linux novice, I was unable to make progress with it. You might gain value from it though.
 
@ypercubeᵀᴹ You've lost me. I was responding to the message I linked back to.
 
@PaulWhite Sorry. my mistake
 
@ypercubeᵀᴹ No worries. (and I see the context message now). The multiple cooperating scans thing.
 
3:33 PM
@Forrest Redhat is different than Debian/Ubuntu. No chance you can have access to a Debian or Ubuntu machine?
 
@ypercubeᵀᴹ Not in the virtual lab
 
Hm. I have a hat, but not a red one. Can I use that?
 
@ypercubeᵀᴹ Oooh, maybe, I'll load up the virtual lab and try again. Thanks!
 
merry-go-round scans!
 
3:45 PM
Yeah. Something I've never been able to reproduce.
 
We haven't been able to reproduce either
it's a nice sounding idea though
 
soo the internet has no copies of SQL-86
this makes me sad.
0
A: Where can I find the first standardization of SQL, SQL-86?

Evan CarrollI reached out to them Hello Evan, Thanks for reaching out. Unfortunately, we do not sell or offer the 1986 version of the SQL standard (ANSI X3.135-1986) for historical reasons. In general, we offer few historical documents, as the revisions mark important changes. However, as your resea...

 
@EvanCarroll thanks for posting it as an answer
 
I just got it back today =(
 
Did a new test of one of our workloads yesterday. Found that scalar UDFs are executed over 2 billion times over a two hour period. Almost all executions are unnecessary.
 
3:57 PM
I was half expecting no response
 
@Forrest of course you can play with new SQL language features on dbfiddle: dbfiddle.uk/…
I don't think there are many though
 
@EvanCarroll at least now we know for sure
 
-1
Q: SSIS - need help on processing data

Ras KnorrI am new to SSIS. I have been given a task to load data from flat files to Oracle Target table. Some how i was able to do this. But i have another requirement to check the flat file data for existence before inserting to the target. Can you please throw some idea(s) which transformation to ...

4 bullets needed s.v.p.
 
@JoeObbish Cool, thanks. I'm personally focused on getting it up and running on Linux. Which of course requires a lot of learning about Linux.
 
4:11 PM
@EvanCarroll and the link I've added?
 
@McNets you mean your answer to the SQL-86 question?
 
@EvanCarroll yes
@ypercubeᵀᴹ Ok tanks anyway.
 
none of that is going to help me track down specifically SQL-86 which is what I was looking for.
I specifically want the first version.
 
@JackDouglas ok, I thought
 
You mention ISO-87 and ANSI-89 but, I want ANSI-86 (v1)
 
4:15 PM
@JackDouglas ok, I thought I can get it with select null
Iso 87 is an alias for ISO 86
 
Yea, but I'm looking for ANSI-86.
The ISO spec is ISO 9075:1987. The ANSI spec is ANSI X3.135-1986. Do you believe that they're the same, if so why?
 
Ok
 
I just want to be able to point out ways MySQL 4.0 < SQL-86
 
You might consult your local library/university library
 
yeah, I bet there are printed copies floating around
 
4:19 PM
I checked ebay, no joy =(
I wonder how many copies they printed?
I would guess fewer than 10k total. It'd be a total collectors item to grab one cheap would be great
 
there are probably easier ways of bashing MySQL though
2
 
@EvanCarroll I've found cross-references between all these specifications
 
@JackDouglas I will bring trivia, nuance, and sophistication.
 
as always :)
 
@ypercubeᵀᴹ The link was exactly what I needed - I was able to install vNext on Linux and run my first query! Thanks cube!
4
 
4:34 PM
Even if you install one SQL server on Linux you must pay for client CALs is it?
 
@McNets SQL Server v.Next is in preview. No licences needed.
 
@PaulWhite ok. Isn't there a date for a prod. version?
 
Not yet. At least, not that I've heard.
There are always rumours of course.
 
Eventually they'll change the name from vNext to vSoon™
 
vAlmostButNotYet RC
 
4:40 PM
@McNets wait a few days ;)
(half joking. MS has a history - in the last years - of making important announcements around April 1st)
 
@ypercubeᵀᴹ Yes
 
My coworker is SUPER confused by OUTPUT clauses. And says it's scary.
I'm so disappointed by this... argh.
 
@ErikE one of the greatest features ever developed, in my book.
 
JNK
@MaxVernon +999999. OUTPUT has helped me stop so many deadlocks
 
4:55 PM
I thought there was/is an issue with RHEL on Azure though. While the SQL Server is free, the RHEL isn't free so people were getting dinged for it
 
@JNK yep, stops 'em dead in their tracks
;-)
 
JNK
well if you have a procedure adding a record then a child record it helps a ton
 
@billinkc if you sign up on developers.redhat.com you can get it for free.
 
ah
 
@JNK where i first learned about it: dba.stackexchange.com/a/36773/10832
 
5:00 PM
output clauses are scary
forces a serial zone
I think that was finally documented in an out of the way place recently
 
interesting
 
for returning results to the client I can accept it. It's not like we have N SSMS results grids for N DOP
although that would be interesting to see in practice
 
sounds like a Paul White special to me
interesting, no results on this
 
"If the OUTPUT clause is specified in the INSERT…SELECT statement to return results to the client, then parallel plans are disabled in general, including INSERTs. If the OUTPUT…INTO clause is specified to insert into another table, then parallelism is used for the primary table, and not used for the target of the OUTPUT…INTO clause"
 
@JoeObbish thank you!
 
5:13 PM
@MaxVernon Still active? I received an email that I can't renew my actual subscription.
 
Is there an admin here, and can admins see deleted comments?
 
No problem. Only found that quoted bit yesterday when investigating something unrelated
 
@JackDouglas ^
 
@EvanCarroll What's up?
 
@McNets I got mine quite recently; I think it was this month.
 
5:15 PM
@PaulWhite can moderators see deleted comments?
 
@EvanCarroll Yes.
 
@McNets - perhaps you need to go through redhat.com/en/developers
 
@PaulWhite can you tell me who deleted the comments on this question dba.stackexchange.com/a/168599/2639. That's really messed up. Someone deleted positive feedback for my answer from the questioner, and also my comment to him telling him how to use the site (upvoting and marking the answer as chosen). That's absolutely unacceptable.
 
@MaxVernon Well parallelism had always been limited to the reading side of the plan, before parallel SELECT INTO etc. came along.
 
@PaulWhite and that part still can be parallel, if I'm understanding that article correctly.
 
5:18 PM
@EvanCarroll It was @JackDouglas. Hang on, I'll look at it.
@MaxVernon Yes, that's my understanding.
 
Thanks a ton. That's not cool at all. I work my butt off for some of these answers. I know I spend hours a day on this site. Removing compliments from the questioners and censoring a request to mark the answer as chosen when the conversation is resolved is pretty over-the-top.
 
@EvanCarroll I don't agree actually. Comments are for clarification only. request for an accept or an upvote will pretty quickly get removed.
 
@Evan There's nothing of value there. I would have deleted them too.
 
I'm looking for the meta, but I thought they addressed comments asking for accept/upvote but my binglefu is weak
 
5:21 PM
Of course there is something of value there.
 
if you consider "thanks" valuable.
 
A pat on the back feels good to you but does it benefit future readers?
 
I want to do a parallel insert into the target table and into the table populated with OUTPUT!
 
The whole purpose of upvotes and marking answers of chosen is negated if we allow users to mistakenly give feedback in comments. That question gets shown as unresolved, and Community bumps it to the top periodically.
 
@billinkc Yeah they can be deleted on just a user flag.
 
5:22 PM
What a nugget [this](https://blogs.msdn.microsoft.com/sqlcat/2016/07/21/real-world-parallel-insert-what-else-you-need-to-know/) is:
>Also, please refer to The Data Loading Performance Guide which has more considerations on ‘Bulk Loading with Indexes in Place’.
 
@JoeObbish Coming on April 1 :)
 
That Data Loading Performance Guide is a nice bit of light reading for a Saturday night.
 
@EvanCarroll No only questions with no upvoted answers are eligible for the Community bump.
 
It's obnoxious, telling a user to mark an answer as chosen when the discourse is over is perfectly fine there is no rule against that.
 
It's also rather obnoxious to tell people how to spend their unicorn points
 
5:24 PM
@Evan Also, the OP currently has insufficient rep to upvote.
 
@MaxVernon It's a classic! Both in terms that it's good and that it's out of date
 
So then don't listen, it's rather obnoxious to run away from a question when the answer goes above and beyond and not to mark it as chosen.
 
@billinkc spending implies that you lose something, no?
 
@MaxVernon I got mine one year ago.
 
@PaulWhite Is it that unreasonable of a request?
 
5:25 PM
@EvanCarroll It can be helpful to a new user (see my recent meta answer) but your comments weren't obviously aimed at helping the OP.
 
@JoeObbish I love some of the content. When I was at Summit 2015, one of the speakers asked how many people had heard of it, then how many people had started reading it, and then how many people had actually finished reading it. I think there was about 3 people left standing.
@McNets oh, I see. you want to renew your key?
 
@MaxVernon and one them was the writer?
 
@MaxVernon yes
 
@JoeObbish Many requests are reasonable. That's one, yes.
 
@ypercubeᵀᴹ undoubtedly
 
5:26 PM
@PaulWhite I'd certainly say he's a new user.
 
What about nested OUTPUTs? I want to keep going
 
@EvanCarroll Yes he is. It's the tone of your comments that makes them noisy in my view. A short note and link to e.g. What should I do when someone answers my question? would be better.
@JoeObbish Composable DML could use some love for sure.
 
Didn't even know that was there. I'll use that then.
 
Fighting the markdown :(
 
@JoeObbish Yes, absolutely, why can't an OUTPUT clause be wrapped in a CTE and used as a source in the next CTE or query?
 
5:30 PM
@ErikE that I would love to have.
 
I suppose that order of operations could get tricky when the optimizer wants to be free to reorder things as it sees fit.
 
298
A: Add data.SE style "magic links" to comments

balphaMost of these now work. My comment below has the following markdown source: On [main], you are expected to write proper English (as advertised on [english.se]), but here on [meta.se] it's more important to have freehand circles, so please [edit] your post, otherwise I'll have to flag you (see...

@Evan Those ^ can be helpful too. I use [help] and [edit] all the time.
 
Like, what if you in a later CTE you insert to a table that you've already queried such that if the query is re-executed it gets different results?
 
Sure infinite loops with OUTPUT would be good too
 
@ErikE "I'm sorry, Dave, I can't do that."
 
5:32 PM
SQL Server usually performs CTE queries once for each reference instead of materializing the results, so it would have to be a major change to the engine to ensure the CTE results are sort of memo-ized.
 
@ErikE That's what I meant by composable DML just above. There is limited support for INSERT.
 
@ErikE Have a demo for a basic CTE that gets materialized?
 
@PaulWhite Yes, I saw your statement and that's what made me say what I did. :)
 
I can only remember seeing it for GROUPING SETS and recursive queries (I think)
 
@JoeObbish Maybe I misused the word materialized. I mean, perform the query once instead of once for each later reference to the CTE.
 
5:33 PM
@ErikE Ah, sorry. I am multitasking (poorly) :-/
 
@ErikE You mean putting it into a table spool, right?
 
@JoeObbish Yes, it has to either be an eager spool or a lazy spool where both consumers can accept the row at the same time.
Except rows are demand-based, so I can't see how a lazy spool could satisfy two demands that may not occur at the same time.
 
If only temporary tables existed.
 
Unless both are waiting?
 
I think we're on the same page. curious to see a simple demo of it. I've never seen one as far as I can remember
@PaulWhite There are poor developers everywhere unable to use temp tables
 
5:36 PM
@PaulWhite were you using OUTPUT?
 
@JoeObbish They lack the skills?
Or a # key?
 
@JoeObbish Please clarify. Do you mean you've never seen a CTE that is referenced more than once, in SQL Server, not result in performing the CTE's query each time?
 
@ErikE No, never seen a simple CTE be put into a table spool
 
@JoeObbish Ok. I've seen it in Oracle, but not that I know of for sure in SQL Server. I think it's a great failing.
 
@PaulWhite They lack the privilege. Could be because of an ORM, reporting tool limitations, etc
There's also a component to it of "I'm not smart enough to figure out when this should be materialized and want the optimizer to decide"
and technically the optimizer does decide for you in SQL Server. It never does it :)
Yes Oracle does it but the heuristic is just terrible
It seems like the heuristics for these things in general seem to be bad in both sql server and oracle
 
5:38 PM
@JoeObbish Explain? I was actually impressed. I don't work with Oracle a lot but when I saw the CTE be spooled and be used more than once in the query it was cool.
 
At one time in SQL server you would get a table spool for a GROUPING SETS query on A JOIN B when B was empty
But no spool when just on A
@ErikE If I remember correctly the heuristic is very close to "if this CTE is used more than one then materialize it"
 
@JoeObbish Could you suggest some of the pitfalls of that? (Perhaps if the data is so large that the spill is worse for performance than doing it twice?)
 
the materialized table has no indexes and no stats (as far as I know)
 
Linked from:
12
Q: Create a plan guide to cache (lazy spool) CTE result

孔夫子I normally create plan guides by first constructing a query that uses the correct plan, and copying it across to the similar query that doesn't. However, that is sometimes tricky, especially if the query is not exactly the same. What is the right way of creating plan guides from scratch? SQLKi...

 
@JoeObbish Oh, I see. It can't properly reorder the CTE parts and do stuff like predicate pushdown because predicates could be different for the two uses.
 
5:41 PM
@ErikE Yes, that's one problem
sometimes you also get the materialized table on the inner side of a nested loop join
 
@JoeObbish And the underlying metadata (as you say, indexes, statistics) is gone after the composed result no longer has the parts and pieces of the original tables.
 
@PaulWhite We have different ideas of "simple" ;)
 
If they ever implement those Connect items, I would expect a sort of hidden temp table with stats, and a feedback mechanism to abandon the materialization if it was counter-productive. And vice versa.
 
@ErikE Yeah. Though to be clear it can definitely be very beneficial sometimes. And there are undocumented hints to force or prohibit it
 
@JoeObbish I've only seen it twice. Once there, and once on some niche performance challenge that I cannot remember enough details of to search.
 
5:44 PM
It's a whole new kind of query algebra that not only has to find the best ordering of operations for a particular set of predicates and joins, but perhaps find the best ordering that can satisfy more than one set of predicates and joins (such that a common branch between them can productively be spooled) then the rest of the query occur.
 
If the memory grant stuff is anything to go by, I would expect adaptive planning to be one of the next big things.
 
Let's say you have a CTE, then when it's used in two different places, one usage has WHERE A = 1 and another has WHERE A = 2, it has to be smart enough to push down a predicate WHERE A IN (1, 2)
 
e.g. nested loops <-> hash join depending on runtime conditions.
 
@PaulWhite If you don't mind I want to make sure I understand your linked article. That guy edited the XML to add a spool to the plan by hand? And forced it with a plan guide?
oh I see your comment now, nevermind
"This is not the most elegant way, but it's a working solution to prevent a costly CTE from being reevaluated." <- I want to know the most elegant way!
@PaulWhite Oracle has it but I personally don't like it (for whatever that means)
We've had some issues caused by it
I can understand why... it seems like a feature that would be very difficult for them to program
 
@JoeObbish Thus why Paul said "if only temp tables existed"
 
5:49 PM
@JoeObbish That post's from 8 years ago, I may have archived some of the details ;)
 
@ErikE No, was responding to adaptive plan comment
 
@JoeObbish There's been lots of research actually it's not so hard (relatively).
 
you shattered the illusion for me :(
 
@PaulWhite that's why he said your posts were kinda old ;)
 
now I feel entitled to it
and want it yesterday
 
5:51 PM
@JoeObbish Not so hard for people like Jim Gray :)
@Lamak Ha! Well I didn't write that one so ner!
 
yeah, but now he can say that your posts and comments are old
 
@Lamak Your mom's old.
 
I bet his dad is too
 
A maternal insult (also referred to as a "yo mama" joke) is a reference to a person's mother through the use of phrases such as "your mother" or other regional variants, frequently used to insult the target by way of their mother. Used as an insult, "your mother ..." preys on widespread sentiments of filial piety, making the insult particularly and globally offensive. "Your mother" can be combined with most types of insults, although suggestions of promiscuity are particularly common. Insults based on obesity, incest, age, race, poverty, poor hygiene, unattractiveness, or stupidity may also be...
There's a Wiki page for everything!
 
@JoeObbish they are
:)
 
5:54 PM
@JoeObbish Me too. Sort of. People will have unreasonable expectations though.
 
@JoeObbish hey, you were the one that said @PaulWhite only blogs about old stuff
 
(paraphrasing)
 
Most query problems can be solved by improving the developer.
 
@Lamak Not what I said!
 
5:54 PM
@PaulWhite love to see that in a DMV ... select sql_handle, plan_handle ,sum(abandoned_materializations) as [ITS A TRAP!] from sys.dm_exec_cached_plans
 
28 secs ago, by Lamak
(paraphrasing)
 
I said that I wanted to write about something new but Paul seemingly has already blogged about everything
 
@JoeObbish Don't mind @Lamak. He's been troublesome ever since he joined a mailing list.
 
LEXMARK_MVP
 
Was it the Most Vain Professionals list?
 
5:55 PM
@JoeObbish That's the fellah.
 
Vein, Vane
 
@PaulWhite hey!...it was way before that
 
@Lamak Ah true. I stand corrected.
 
now, that's better
 
@swasheck Best DMV column name ever.
 
5:57 PM
@EvanCarroll have a look at the answer: verycomputer.com/165_1797958070ed8752_1.htm
3
 
That's an awesome contribution!
 
An off by 3 error.
 
@McNets you should update your answer with that, that's a great find.
 
00:00 - 18:0018:00 - 00:00

« first day (2277 days earlier)      last day (2582 days later) »