The Heap™ – Consultancy ©®

is there a common nomenclature for how different platforms handle null interpolation to/from csv?

also, morning

like... i'm pretty sure "NULL" just isn't handled in CSV (at least not in the RFCs i've read), but obviously all the major platforms have their own idiosyncratic ways of dealing with it

i'm wondering if there's a good way to talk about that handling in general while trying to wrap my head around a problem

John aka hot2use

6:52 AM

Morning

8:25 AM

Morning

8:42 AM

morning

ypercubeᵀᴹ

morning

Tom V - try topanswers.xyz

8:59 AM

@PeterVandivier should "The Empty String" be the name of the bar I open when I can't take IT anymore? 🤔

3

@PeterVandivier Might sound a bit weird for non IT folks

i'm sure i'll be drunk enough that that doesn't bother me

10:28 AM

@PeterVandivier I hear you. The app I'm working with just now allows nulls for all types except string, which are held as empty strings. Oh, and did I mention all data is passed to stored procedures as JSON? Hash tag keeping me employed!

@MichaelGreen cheezus .. why json? they could make it even worse.. xml :d

11:02 AM

@Marian It's a "low code" application development environment. It has no out-of-the-box support for stored procedures, and even less support for table-valued parameters. And no-one, myself included, wanted to deal with @product1, @product2, @product3 ... It does, however, have a generic ToJson() function which works on any of its internal representations.

Meh, I got to learn a lot about JSON parsing on someone else's dime. Anyone looking for a Data Engineer??? : - )

@PaulWhite9 Um .. by any chance, have you bought a Land Rover recently??

bbc.com/news/world-52108765

Ha ha no but we did put some bears in a shop window

3

11:18 AM

Nice. Over the years a small corner of our front garden has accumulated some gnomes. One morning last week we woke to see they'd acquired a sign from "The Department of Gnome Health" and were spread along the property line, 1.5m apart.

3

😁😁😁

Tom V - try topanswers.xyz

Also, I love the nominative determinism in that article:

> “It’s just fun,” said Ms Ha

John aka hot2use

😂😂😂

12:02 PM

@MichaelGreen hahaha, that's brilliant

0

Q: How do I make our data warehouse OSHA compliant?

What are the safety techniques that we should follow in our data warehouse to guarantee OSHA compliance? The existing literature isn't very helpful. I bought yellow high visibility vests for all of my DBAs but they're refusing to put them on.

> I bought yellow high visibility vests for all of my DBAs but they're refusing to put them on.

😂

Colin 't Hart

Just saw that. April 1?

Prashin Jeevaganth

Does anyone know any way to store values into 4 declared variables in a procedure using a single PostgreSQL query? I had to duplicate my queries 3 times to make mine work paste.ofcode.org/Y2iSdWdkZTAjm3csdAGEkw

1:02 PM

@TomV-trytopanswers.xyz to wear inhouse? :)

1:48 PM

@Taryn That slow AG metadata query: Do you have an execution plan for it? Does a hint like OPTION (HASH JOIN) help, or is no plan possible?

I would try it but I don't have access to any systems with an AG right now.

@PaulWhite9 it's part of Ola's backup script, let me get a snippet and the plan

Oh I see, thanks.

it's the underlying HADR CLUSTER API call that appears to be the issue

we're seeing high waits for that

Right that makes sense.

but it of course wasn't an issue in 2017

1:49 PM

That makes less sense.

Hey perhaps it'll be fixed in CU17

yeah, probably - I can get you the plan shortly - it take about 4 minutes to run the query on our secondary - that makes taking backups super slow all of a sudden

I bet

it was taking about 3-4 hours, we're up to 22 hours

twitter.com/tarynpivots/status/1245345392796950529 for others wanting context

and this - twitter.com/tarynpivots/status/1245348385118437376

2:00 PM

@Taryn I wonder if this would be any better:

select db.name as [ReplicaDBName], ag.name as [ReplicaGroup], ars.[role] as [ReplicaRole]
from master.sys.availability_groups ag
inner join master.sys.availability_replicas ar on ar.group_id = ag.group_id
inner hash join master.sys.dm_hadr_availability_replica_states ars on ars.replica_id = ar.replica_id
inner join sys.databases db on db.replica_id = ars.replica_id;

Much as I hate to use join hints generally

alright here is the plan - brentozar.com/pastetheplan/?id=S1bpCfzvI this query is directly from Ola's script

i mean... i'm gonna assume caching isn't an option 🤷‍♂️? tjaddison.com/blog/2017/10/…

running that on our remote secondary with 370 databases takes 3.5 minutes

at least that's how we bypassed hadr_clustapi_call trubs back on the farm

but i'll just go back to the kids table now

that would work too, yes. bit more work than adding a hint

not that the hint is super likely to work

2:10 PM

it did - went from elapsed time = 210281 ms. to elapsed time = 2309 ms.

Lucky break. That should probably be in Ola's script until Microsoft fix the underlying issue.

I'm going to open a bug with him :|

feel free to share the hinted query in case it is helpful

thanks for the details

thanks I will. We also are suggesting a correlated subquery, which is faster as well. I'll credit you for the hash join :)

Oh I didn't realise you already had a fix

2:35 PM

@Taryn did Kin's solution work? dump those sys views in temp tables and query those

@Marian not going that route since it would involve a significant change to Ola's backup script

@Taryn oh, I see

Andriy M

3:15 PM

There's more than one way to open a bug (and it's not always a very pretty sight).

Q: Is a display name sufficient disclosure?

-2

I just raised a flag for "moderator attention" on this answer with the following message This answer from a Microsoft employee recommends Azure without any kind of disclosure of conflict of interest in it. I do not consider the user profile and bio to be sufficient disclosure here because the...

discussion moderation profile-page

I didn't handle that flag by the way, but I would have declined it as well.

In fact I would have declined it twice.

mustaccio

3:33 PM

@PaulWhite9 Technically he's right though

I'm not saying that particular answer deserved a flag

4:17 PM

@mustaccio is he though? In the case of that question I’m more than happy to say “yes, display name disclosure is sufficient”

4:36 PM

@PeterVandivier same for me..

and to be fair I voted that answer as being way better than the OP's own test and answer (even though I could be biased and say I'm pals with the OP)

the proposed answer was a bit better than the "We can't afford currently a real test scenario of restoring the 2 TB backup and I don't know yet the final configuration for the migration" .. this is not how it's supposed to work for production

Andriy M

I had a similar discussion with Sean once, though I can't seem to find it any more. Maybe it was in comments under one of his answers, or I just can't find the right keywords to find it in chat. I think it was about something other than promotion, though. Something to do with what he was suggesting being authoritative enough or something like that.

But I remember suggesting he disclose in the answer his affiliation with Microsoft in that specific case

IMO disclosure is (in part) meant to support an answer of ambiguous quality. You want to be transparent if you’re speaking on a niche topic where your advice could be seen to benefit you at the expense of the asker. But if we don’t require Brent O to add 100 words to every sp_blitz answer he’s ever given, then a vague framing about “use a VM, azure is an option” is fiiiiiiine

disclosure prevents extraction of value from askers at the expense of effective advice

ypercubeᵀᴹ

5:03 PM

bad bot

@ypercubeᵀᴹ slow bot :)

Q: How will Coronavirus affect Social Security solvency?

0

The Coronavirus (known by younger millennials as the "Boomer Plague") is estimated to have a cost of life between 100,000 - 240,000 people. According to the CDC approximately 80% of deaths occurred among adults aged ≥60 years If we're to assume those numbers, 80,000 - 192,000 people over th...

social-security

5:19 PM

@EvanCarroll perhaps you should review your question. theguardian.com/us-news/2020/apr/01/…

How does that have anything to do with the quesiton

> approximately 80% of deaths occurred among adults aged ≥60 years

Yes, and from your article,

> The big difference between them and older adults is [young adults’] mortality rates tend to be lower.

So we may go the hospital, but certainly even that is substantially less likely.

But we almost certainly won't die from it.

Sooo. the question seems to be totally valid, no?

@EvanCarroll of course

first 3 deaths in my village was < 60

That's amazingly unlikely

but it can certainly happen.

I mean each death has a 20% chance of being UNDER 60.

5:23 PM

suggesting is ok, sure. But I don't see it as necessary. I can read more than the stated answer's listed words. Their meaning has weight because they come from a valued member of the community.

And here we usually establish that by reading beyond the answer - we check the answer, the answerer biography, maybe the other answers..

@EvanCarroll this are the numbers

So the chance of having three deaths under 60 in a row, are 0.20*0.20*0.20, or 0.008, ie 0.8%

no, if you're looking for a non important reason :)
https://www.grammarly.com/blog/affect-vs-effect/

but it's still possible.

@Marian If I'm talking about the Coronavirus doing something to social security solvency, that something it's doing is a verb and thus "affect" right?

right, I was just trolling.. sorry, don't know what SS solvency is.. :)

5:27 PM

Social Security Solvency is the problem that boomers created when they forced upon us a system that required us to pay for it, while giving us no chance to benefit from it.

when that retirement "fund" runs out is the period of insolvency.

Essentially a group of old fuckers decided that if they make more money than $130,000 they shouldn't have to pay into that fund. So their generation's super-wealthy don't get the full burden of the tax and thus they're net-takers and the fund is insolvent losing money as they retire and draw down it -- all the while the young pay for it.

Cade Roux

@EvanCarroll It's actually does not seem unlike at the start of an outbreak. The younger deaths have been cardiac and not respiratory. They are sudden and so those people have not been hospitalized for long time and succumbing.

Perhaps the profile changed. Not sure, we're also a fat ass nation too unlike China. So we'll probably die more and such.

I blame you for that.

Cajun food is good shit.

Damn Louisiana couldn't invent a low fat jumbo or jambalaya. Now everyone in the south is going to die.

@EvanCarroll thanks, sounds a bit more complicated than the standard European pension.. :)

I think Europe is in a far better situation. Their boomers are far less sociopathic and right-leaning than America's boomers. Even the UK has some ratio of Skinner:Thatcher that is non-Zero.