The Heap™ – Consultancy ©®

6:32 AM

Morning

Morning

Morning

7:41 AM

Morning

Morning

7:54 AM

mornming

I was thinking what a good title would be for a blog post about the recent SO events / fiasco.

Stack Overflow vs users: What a Travesty!

too provocative?

@ypercubeᵀᴹ Not provocative enough for me

I suppose it's hard not to be provocative when you are feeling strongly about something

I think it's correct to be provocative in some situations

For me personally I'd like to reflect the provocativity of the SE staff.

Generally, I consider a provocative title to be a good idea if it's backed up by good points in the article body. Or if the goal is to open a discussion, I guess (if the blog has a comment section).

I mean, it would probably be disappointing to click a link only to encounter generic whining, without being able to tell the author how pathetic their click-baiting title was.

9:22 AM

Burninate stored-functions?

Trans-gendering Stack-Overflow

Q: Solution for the deletion of old Database Backups for Linux SQL Server 2017 after Backups created with Ola Hallengren scripts?

9:50 AM

0

I have a problem when I want to delete older backups, I've created with Ola Halengreen scripts. USE Maintenance EXECUTE dbo.DatabaseBackup @Databases = 'USER_DATABASES', @Directory ='/mssql/backup/', @DirectoryStructure ='${InstanceName}{DirectorySeparator}{DatabaseName}', @Filename='{Databa...

Michael Green

11:47 AM

@George.Palacios At my last place we used that to host the bots. It had its problems but was increasingly good. Last I heard they'd got some investment from somewhere or other so expect continuing improvement.

Michael Green

12:04 PM

@ConcernedOfTunbridgeWells @CarefreeOfCanterbury? @InsouciantOfIndonesia?

3

dezso

12:34 PM

meanwhile in Milan

12:48 PM

@dezso That's in front of the Borsa right?

1:11 PM

i took the liberty of burning stored-functions from q/234524 & q/251195

seemed trivial enough of a change but afaik, i'm too low-rep to actually burn the tag itself

An unused tag self-deletes

👍

🤔do low-rep users have ad-hoc access to deleted content via SEDE?

i assume there's situations where a deleted post has bogus tags left on

@PeterVandivier rep doesn't matter on SEDE, so no deleted content isn't there

@PeterVandivier What's SEDE?

@ypercubeᵀᴹ was I even making sense here given it's mysql?

You cannot just replace "where" with "having" and expect the query behaviour to be the same. The evaluation order is different and the query might end up being a lot slower. — Tom V 6 hours ago

1:18 PM

@George.Palacios data.stackexchange.com

@PeterVandivier Ah cool, thanks

Would a substring in the where even consider an index?

@TomV i assume the same data gets scrubbed out of the "official" dumps as well?

@PeterVandivier You only get to see what's public

presumably if a mod or high-rep user wanted to perform such a check "Find low-frequency tags that only exist on deleted posts", they'd be limited to the public search endpoint?

i'm guessing there's no aggressive background process for such edge-case cleanup

@George.Palacios Stack Exchange Data Explorer by way of explaining the acronym as-used :)

dezso

1:26 PM

@TomV it is

@TomV yes you are right.

For the specific condition, I don't think it would make any difference probably though.

1:51 PM

@ypercubeᵀᴹ I wanted to make clear to him that he shouldn't just suggest it as an alternative without any side effects, but I didn't feel like going into an argument about it

He's not having your remarks either :)

2:26 PM

What is this mess :o dbfiddle.uk/…

@TomV MessSQL

3

@TomV As I understand, having <condition without using aggregate functions> works the same regardless of the sql_mode state, i.e. as a synonym of where (provided there's no group by, that is)

Cade Roux

Does that excuse it from getting the answers wrong?

That was actually more a response to your comment on that answer @TomV

Yes but I can't figure out why in this the where 1 returns a row and the where 5 doesn't return 2 rows

unless having behaves the same as where when there is no group by, but is allowed to return random crap when you have a group by and use having on a field that isn't part of the grouping

2:34 PM

That was exactly my question to the answerer

And the point of my comment. I disagree with the "you can just use having instead" without any explanation

Although I would expect one row, not two

I understand we have to be welcoming and I haven't downvoted his answer but instead tried to guide him into clarifying some of the side effects and prerequisites. Instead he decided to argue there are none but I still disagree

My guess is, having behaves like where as long as there's no group by and no aggregate functions. As long as you include an aggregate function, having becomes a normal having

But I don't know, hence asking him (them)

My guess also doesn't explain the absence of rows in the scenario we've just discussed

The answer should at least have a warning "only use this when not actually aggregating, otherwise no-one can explain what MySQL considers the correct resultset"

2:52 PM

Can anyone ever explain what MySQL considers the correct resultset?

In spite of everything, I still hope someone can

3:16 PM

Where there is life, there is hope

@TomV You used set sql_mode =''; That reverts the behaviour to version 5.6, regarding GROUP BY.

@AndriyM about the absence of rows in some queries, this may help: dbfiddle.uk/…

@ypercubeᵀᴹ Ah of course, the column is not aggregated so the value is undefined

Thanks

In this case it just happens to be 1

The tricky (to figure out) part is that if there is an aggregate (either in HAVING or SELECT), the query acts as if there is a GROUP BY. This is fine and good and expected. But allowing to refer to non-aggregated columns (HAVING b=5) produces this nonsense

I guess internally MySQL may still follow some rule but of course there's no point figuring it out if it's undocumented

And what the answer that Tom linked suggests something else. Using HAVING without no aggregate functions anywhere. That (non-standard) results in HAVING acting like WHERE.
But you allowed to refer to SELECT aliases in HAVING.

So, it kind of evaluates in the order: WHERE -> SELECT -> HAVING

3:26 PM

51 mins ago, by Andriy M

My guess is, having behaves like where as long as there's no group by and no aggregate functions. As long as you include an aggregate function, having becomes a normal having

So my guess was right, I understand

It's good that they fixed this in versions 5.7 and 8.0 (and adding window functions required that fix). I guess we'll need some 20+ more years for all mysql users and devs to get used to it.

@AndriyM yes

3:59 PM

@PaulWhite Where there is life, trans-uranics may not be used.

@ConcernedOfTunbridgeWells I understood that reference

@PaulWhite Fabulous.

I have them all on DVD

We rewatch them from time to time

@PaulWhite That's some serious nerd cred.

Amazon supplies all

4:05 PM

It's been quite a while since I've seen a Sapphire and Steel.

Some of the episodes still freak me out

Not as much as when I was a kid

But still

The nerdiest thing I did recently was to find a Dalek in an antique shop just around the corner when I was still living in Sunningdale.

The label at the bottom says 'Original 1963 Dalek prop' but I'm reliably informed that it's actually from the Peter Cushing movies of a few years later.

Somebody on a forum knew enough about Daleks to recognise the specific version and post a correction.
I was comprehensively out-nerded.

Hello, is there a good way to display tabular data in questions here? I could have sworn there used to be. . .

@peacedog No. There has been pressure put on the SE developers to support tabular data in questions but they haven't ever done it, despite having implemented it on other boards.

I can't believe they didn't implement one of the extended markdown libraries just to get tables. Alas

I'll figure out a way to get it displayed properly for this question, thanks!

Q: Please can we have markdown tables on dba.se

4:12 PM

67

Several sites have MathJax enabled, so there is obviously some room for site-specific markdown. Therefore please can we have markdown tables here on dba.se? There is an open feature-request for this on mSO, but whereas it is hardly a key feature for a programming Q&A site, a significant percent...

feature-request status-declined

3

billinkc

Or just load the data into a fiddle site and we can use proper tools

2

Max Vernon

4:53 PM

@peacedog ozh.github.io/ascii-tables

@MaxVernon ooof, I didn't realize I needed to format the ascii as code. Icouldn't get that working before. Thanks.

Max Vernon

@peacedog cool

Q: Grouping Data Based on a Header Record

5:30 PM

I am not really sure if I have titled my question very well, sadly

0

I have a spreadsheet I'm importing into the database raw. It has records in what is known as the "Lockbox format", used by banks apparently. It looks something like this generically: Batch1 details child record another child record summary for batch Header Row for Batch2 I have a fiddle set u...

sql-server t-sql

8:01 PM

@ypercubeᵀᴹ I've read 'sql Server'