The Heap™ – Consultancy ©®

8:29 AM

Morning

10:08 AM

ƃuᴉuɹoɯ

10:24 AM

Somebody predicted a while back that the all-new™ improved™ Stack Overflow Executive Leadership®, after having successfully fired/alienated most of the people holding the place's institutional knowledge, will, over the coming years, step by step, incredulously rediscover the things that Jeff & Co. had already discovered and built into the way the site works back in 2009. Looks like we're at step "having a Meta is in fact important even though it is really really annoying to have one" now 😄 — Pekka 19 hours ago

5

😀😆😀

10:36 AM

Is the "Users are really annoying, let's push a load of them away" step complete? ✅

😂

Q: Why *not* ERROR: index row size xxxx exceeds maximum 2712 for index "foo"?

11:38 AM

3

We have repeatedly seen failing attempts to index columns with values exceeding a maximum size. Postgres 10 has this kind of error message for it: ERROR: index row size xxxx exceeds maximum 2712 for index "foo_idx" HINT: Values larger than 1/3 of a buffer page cannot be indexed. Consid...

underrated post

Let's derrate it then

rite tho‽

apparently that compression applies to jsonb as well dbfiddle.uk/…

i just had to walk back a whole 15 minute talk i gave to a dev about maximum key size

oh well... back to actual work i guess, that was an informative hour and a half 🐇🕳

uh oh, looks like the rabbit hole is continuing

dammit!

James

1:25 PM

@mustaccio no for update - we've got this mysql DB that takes the vital signs of patients, 1 row, per minute per device. Day 1 and it's past 1GB. My SSIS package runs select *'s where timestamp >= yadda yadda to put the data into SQL server, but i'm worried that those select *'s are going to prevent inserts

@James not so much assuming you have a good indexing strategy

James

1:49 PM

hmm, the mysql might not - it's a vendor db

i went and spamed 'set transaction level read uncommited' throughout the etl package

but I think I also have to talk to the bossman, it looks like it might be 1 row per second...

Mathematics

3:19 PM

Heap is on !

billinkc

3:34 PM

H * eap is on!

H * e * ap is on!

Glenn Frey, The Heat Is On

►

3:51 PM

god damn how is this question off topic, stackoverflow.com/q/60302874/124486

what in the hell is wrong with these people

@EvanCarroll some people hey! oh, wait, it's Stack Overflow

Anybody notice the q in my early morning post?

@Johnakahot2use :-)

@EvanCarroll it's open

lol, meta.stackexchange.com/questions/343890/…

> To all of the moderators who have resigned or suspended your activities over the past few months: your presence and impact is missed. We value all of your work to keep your sites clean and communities healthy.

its a bit goddammed late for that, isn't it?

4:17 PM

appreciate that. i just don't get these people

Q: The company’s commitment to rebuilding the relationship with you, our community

27

My name is Teresa Dietrich. About a month ago, I joined Stack Overflow as Head of Product and Community. During my years as an engineer and technology leader I saw the impact this community and site has on people across the world and I am very excited to be here. While I have personally gotten a ...

discussion featured community policy announcements

It's a start.

It’s still ridiculous that they fired a few long-standing community-based employees & hired a new one. Making people redundant & then filling their shoes straight away is illegal in the UK

Got to love US labour laws

@Johnakahot2use Not at the time, no. I see it now.

@Johnakahot2use wich q?

@McNets One of the "morning" words is spelled "morninq"

@MaxVernon Yes. They should start by undoing the obvious bad things. More words won't fix anything. Actions are needed. Probably too late anyway.

4:25 PM

@PaulWhite9 ah, it was a kind of "Find the gap"

Well that just reminded me I turned the volume up to watch a video a few hours ago.

Yikes.

@Johnakahot2use Would you like it fixed?

@PaulWhite9 I wonder how negative my answer will go, lol.

I don't have a meta account to vote with.

meta.se is a weird place lol

its gone up and down like a damn yo-yo so far

~~Well thanks to the new and more welcoming 5:1 rep ratio you will be well ahead.~~ Oh that was questions. LOL

4:31 PM

I lol'd when I read the first sentence in her post:
> My name is Teresa Dietrich.

trying to fend off the negativity, I bet.

How so?

by making clear that she's new.

Oh I see.

or at least not anyone who was involved

well, damn, if Shog agrees with me, I don't need anyone else to vote:

I tend to agree with this answer. I like nice words. I'm a big believer in nice words. I try to write nice words... But it is actions that tell you whether nice words are sincere. There is a certain action the company must take which has been pending for several months... It could be done well, or it could end up being another slap in the face for the volunteers who've dedicated so much time here. I will withhold judgement on these words until I see how that action is taken. — Shog9 2 mins ago

Been a while since he dropped in here

4:36 PM

it certainly has

I think you'll find Shog is agreeing with me. 😀

@PaulWhite9 no. It’s fine the way it is. Thank you.

@Philᵀᴹ the US has labour laws?

billinkc

Sure, you can't offer hair braiding services unless you've been to an accredited cosmetology school for 2 years and have passed a licensing exam

@Johnakahot2use I wasn't offering to change it, just wondering what you would like to happen.

Amusing watching the people on meta who are prepared to believe it will all be different this time.

4:44 PM

@PaulWhite9 oh, yes, my bad :-)

@PaulWhite9 very

@PaulWhite9 I want SE to be a better place. When will you deliver?

oh eff me...

What about the damage done to the countless number of victims persecuted and suppressed by your elite class? — Tim 2 mins ago

@Johnakahot2use Probably never, to be fair.

@McNets The silent q as in “Fishqer”

A: The company’s commitment to rebuilding the relationship with you, our community

10

I want to be hopeful. I really do. But I can't. The company has burned through all it's goodwill, with me, at least. I want to think this is the first step to recovery, but I thought that about every communication from SE to the community. I'm not willing to believe in yet more words. Words...

4:49 PM

Closed the tab.

3

wise, very wise.

billinkc

5,4,3,2,1

community wiki
Tim

He has deleted the answer

Let's not give that topic any more oxygen.

5:35 PM

All this talk of Elite reminds me what a good game it was

5

6:19 PM

@PaulWhite9 why should we not give more oxygen. Oxygen is necessary for LIFE, at least for most of the kinds that Jesus cares about.

@Philᵀᴹ that goes back a ways

Jesus had anaerobic friends.

so you want to create a condition in which T can thrive, because you believe he's like one of Jesus's anaerobic friends?

how considerate

@CadeRoux waitwut?

plz EAGL

You can pass comment when you become a member of the Elite, Evan

6:34 PM

Wait, are we making him eligible for that?

I thought we discussed that last week in our secret Elites meeting

shhhh, Cole.

right right

it's not secret if everybody knows.

/jedi wave

this is not the meeting you're looking for

6:36 PM

Wonder if the new community manager will make up more rules that stop you being mod electable 😉

My name is Peter and I'm a DBA

(did i do that right?)

my router has two WiFi APs, NotTheWifieYou'reLookingFor and TheWifiYou'reLookingFor - any guess which one you want to connect to?

It's been 3 days since I last dropped a table in production

@PeterVandivier lovely :-)

@PeterVandivier appplauds

@MaxVernon Is it the secret 3rd option?

6:38 PM

@PeterVandivier no but there should be. Also, my WiFi Pineapple isn't always on

@Philᵀᴹ I mean, I'm already unelectable :p

6:58 PM

in case anyone feels like being super awesome - github.com/microsoft/vscode/issues/89120 <-- if that gets to 20 👍s in the next month it'll get added to the offical backlog 🙏

pretty sure i haven't shilled for that here yet, but i really want it to get looked at

Tom V - try topanswers.xyz

👍🏻

@PeterVandivier 7 to go

Tom V - try topanswers.xyz

:D !!!

I wonder how long it will stay in the backlog after it gets added

6

7:06 PM

@TomV-trytopanswers.xyz lol, mystery-ghost.gif for sure, but knowing it's not "officially declined" would be a step in the right direction imo

idk why, but crazy long unreadable json payloads really trigger me for some reason

and that feature feels like the "tabs-v-spaces" solution to the whole mess. like - i don't need to care if you dump a novel in your json payload if my editor is intelligent enough to interpret the line breaks

@PeterVandivier 5 to go

^_^

party-parrot.gif

3

that's a party parrot

Monty Python: The Parrot Sketch & The Lumberjack Song movie versions HQ

yaas

James

►

7:22 PM

I had that song on CD 🤣

If I hadn't nailed that parrot to its perch, it would've muscled up to the bars, and "voom!".

Mate, that parrot wouldn't "voom" if you put 4,000 volts through it. It's passed on. Gone to meet its maker. It's climbed the curtains and joined the choir invisible. That, sir, is a DEAD PARROT!.

(taken from memory, don't @ me, lol)

8:20 PM

@PeterVandivier 17

8:53 PM

💪

love this chatroom

got as much progress here as i did from twitter/sql-slack/ps-slack/work-slack/personal-gchat combined

y'all are the cat's pajamas

@Taryn - I just realized I have two accounts on stackoverflow under two email addresses. If I ask to get them merged, what happens to them? i.e. will both emails be attached to one account? Does one of the accounts just go away?

@PeterVandivier feel like joining another slack, lol? some of us regulars chat there once in a while

@MaxVernon gotta keep those unicorn points on lock

they're, like, worth a lot.

lol

@MaxVernon you don't even know. i've got 5 that i keep generically open (if mostly muted)

@PeterVandivier lol, invite incoming

9:00 PM

i just only asked in the 2 accounts that have #vscode sub-channels

i have clearly painted myself into a corner

lol

u have

sry, train wifi

i've been wondering for a sec though - what's the "ct" in "mvct"?

"Consulting Tech"?

@MaxVernon they merge them, the second email goes away, rep transfers.

"Canadian Technologist"?

@PeterVandivier have you tried on your whatsup's groups? 😂

I'll try in one of mines, does it require a github account?

9:12 PM

@MaxVernon That's one massive Norwegian Blue

9:31 PM

@AndriyM haha isn't it though!

@McNets yah

Looks massively dead too

Sure is. Very.

9:52 PM

@PeterVandivier 19

10:28 PM

There's no clever way in a SQL Server table to have computed columns depend on each other?

I have a table of test input, test output and expected test results (large nvarchar columns), I have a hash computed on each column, but because some of the tests are generated and can vary as mapping tables evolve, I don't really have a good handle on a particular test besides the set definition - the hash of the input definition is basically as good as that.

The hashes are persisted. So I'd like to use that hash in another computed column of the test "name". Seems like to have the "name" in a persisted way, that I would basically need to make an indexed view on top of the table?

@McNets you guys are the best 😄

Basically:

CREATE TABLE [TEST].[TestHistory] (
[TestHistoryID] [int] IDENTITY(1,1) NOT NULL
,[AscendStudyID] [int] NOT NULL UNIQUE
,[Source] nvarchar(250) NOT NULL
,[TestSet] nvarchar(250) NOT NULL
,[SetDefinition] nvarchar(max) NULL
,[IEVSourceID] int NULL
,[Name] nvarchar(250) NOT NULL
,[TestInput] nvarchar(max) NULL
,[TestInputHash] AS CHECKSUM([TestInput]) PERSISTED
,[ExpectedOutput] nvarchar(max) NULL
,[ExpectedOutputHash] AS CHECKSUM([ExpectedOutput]) PERSISTED
,[TestOutput] nvarchar(max) NULL

That last one is not allowed because it depends on the hash, so the only option is really to repeat the CHECKSUM(TestInput), which probably makes it evaluated twice?

@CadeRoux A regular column and a trigger to populate it?

10:50 PM

@AndriyM I guess that's an option, I can also just populate with a single update any time all the tests are defined.

There's another more complex column that I need defining success or failure and it might make more sense to have it all in a view.

If a view suits you well, it's definitely a better option than a trigger.

In my view anyway, if you pardon the pun

While I'm here, there is a particular SELECT * pattern I've been using a lot within procs, and wondering if there are any problems with it:
WITH somequery AS ( SELECT specific columns and whatever )
, someotherquery AS ( SELECT specific columns and whatever )
SELECT * FROM somequery
UNION ALL
SELECT * FROM someotherquery

If the SELECT * is only within the proc, there's no danger of issues with changes to underlying tables/views/functions, right?

Like everything outside of the proc is always selected explicitly.

It references columns defined right there in the proc. I would be find with that kind of querying in my production code.

But these subqueries I often add and remove columns a lot as I need new parent linkage and things.

It's already enough work changing up the sets (and potentially the output), but not having to also change up the UNION part every time is useful.

I agree

11:04 PM

It hasn't caused me any problems like SELECT * in a view might when the table underneath changes, just curious if there's something I overlooked. I would think that within the proc, all those * are already locked down when it is parsed, since all the preceding tables/columns being pulled in are already fixed too.

I sometimes use SELECT * on a table variable, though in such cases I tend to place it near the TV's declaration. When the declaration is far from where I'm querying the TV, I still prefer listing the columns explicitly. It's not a problem with a CTE, of course. You are defining a column list in a CTE definition and querying the CTE immediately

@CadeRoux I can't think of any potential issues, as long as it's in a proc (as opposed to, say, a view definition)

I have a lot of things like this in the middle of CTE chains:

,dst_not_src AS (
SELECT *
FROM dst
EXCEPT
SELECT *
FROM src
)
,src_not_dst AS (
SELECT *
FROM src
EXCEPT
SELECT *
FROM dst
)

With a long chain of these it may become somewhat difficult to parse, I guess. Depends on how tall the query is, specifically on how far the last SELECT * from the place that defines the actual list. But it's only about readability for me, really, and not about some kind of potential syntactical issue that could unexpectedly bite you when you change your code or something.

I do worry about execution issues. I did have some problems like that way back in SQL Server 2005 and had to automate refresh sql module on some of our objects during updates. That system was always live, though. A test env and production env and when we migrated changes to prod, there were some automated checks we did after the change script ran.

This system is recreating all procs, views and functions on deployments, so probably wouldn't run into it. That system had select * in most of the ITVF function layer that tied all the simensional stars together so that they would not need to be updated all the time.