@ConcernedOfTunbridgeWells It's bad enough with 1, we have 5 and no intentions of moving off that.

@mustaccio procrastinator?

Morning

Morning

Morning

twitter.com/Xof/status/1437559505873559552

A chairde - Morning all!

@Colin'tHart #PostgreSQL? Is there something database related in that passage/quote that I'm missing? Context?

@ypercubeᵀᴹ Christophe Pettus CEO PostgreSQL Experts Inc? I was up late last evening, so maybe I'm a bit slow this morning but I still.l don't get it? It's not only jokes that have to be explained at great length!

@Vérace I don't get the post either. Not without context anyway

Exactly!

I guess it has to do with some grey hacker attack

@ypercubeᵀᴹ Googled him - nothing about grey hat stuff that I could see... PostgreSQL Experts Inc. are reputable AFAIK - Josh Berkus is with them - a major contributor - videos on YouTube re. PostgreSQL &c...

Oh, I thought many of you would have seen this:

postgresql.org/about/news/…

:( yea

Yes, PostgreSQL Exports is a reputable company, and their CEO was commenting on someone else's actions in trademarking Postgres in Europe

@Colin'tHart - I see, so it's an "I'm doing it for the good of the children!" argument on the part of Tortosa and the Fundacion PostgreSQL? Didn't someone try to do that with Linux in the early days?

I may be a "PostgreSQL Ayatollah", but I can't keep up with every PostgreSQL fact and figure... Maybe I should register a company under the name of "PostgreSQL Ayatollahs" and that will scare the PostgreSQL Core Team and the PostgreSQL Community Association of Canada (PGCAC) into giving me the rights, which I will kindly sell back for the "good of the children"?

It's an interesting story mind you - glad you brought it to our notice!

Should have written "so-called Fundacion PostgreSQL"... or, in a spirit of European multi-lingualism, soi-disant?

@Vérace aubit.com/index.php?page=ProductDetails/Aubit4gl

At one point they had an Informix SQL dialect emulator for Firebird, but I'm not sure if that ever made it into the main tree.

It's a real thing - apparently there are quite a few legacy Informix 4GL apps running on it now.

@bbaird And half of them running on crusty old legacy platforms, I bet.

@Colin'tHart I am reading about the dispute

@ConcernedOfTunbridgeWells - wow, first time I've heard of it! There was a commercial project called Fyracle which was meant for Firebird to emulate Oracle... I haven't seen sign nor sight of it for years (Googled - nothing of interest). There was also talk of the embedded server becoming the db (sort of Access) for LibreOffice - came to nothing AFAICS. Pity, I'd like the db to have some sort of flagship project...

@Vérace IIRC Fyracle was a spin-off of Compiere, which is an open-sourced ERP system that was dependent on Oracle at one point. I think it was later modified to support Postgres natively, which obviated the need, so the project fizzled out as it no longer mattered.

There's actually a fairly active ecosystem of consultancy firms doing Compiere implementations, or was at one point. It looks like the outfit that made it has repositioned it as a cloud-based offering.

@ypercubeᵀᴹ - just checked out that link - it's "Álvaro Hernández" - recognised the name. He's an AWS data-(anti-?)-hero behind Ongres and is also behind what I thought was a very interesting project - Babelfish - a project to allow PostgreSQL to emulate SQL Server.

With AWS behind it, I was hopeful that it might really work this time - but it appears to have stalled. Mayby those proposing it (AH?) were unrealistic in their expectations? Vapourware for the moment - I'd be interested in giving it a twirl when it does come out - or maybe I'll be drawing my pension by then...

Yes, I had read some of the discussion about babelfish

@ConcernedOfTunbridgeWells Yeah, that rings a bell - found this last release 2007... which is almost as old as I am! :-) Dead as the proverbial doornail. I have a feeling that Babelfish might go the same way... maybe Jeff and Bill got together and canned the project? Speaking of Bill, how about that for a fall from grace?

@Vérace the original announcement said that they will make it available in github in 2021. We still have 3 months of hope ;)

@Vérace Disappointed but not shocked, I guess. I wonder when we'll see Larry in the news. He used to be famous for dating 19 year olds.

@ypercubeᵀᴹ I guess it was that response that triggered Christophe Pettus's stolen TV analogy.

@ypercubeᵀᴹ - I'm afraid that I'm not that hopeful. If I were a betting man, I'd give you 4/1 against it appearing anytime soon. Ongres' blogs have dried up (only 4 in all of 2021 so far) and they have this (which I didn't see before) StackGres. So, my thoughts are that they've moved on. This Fundacion fiasco strikes me as the desperate flailings of a drowning man - or maybe this is a backdoor attempt by AWS to take the PostgreSQL name? A la Lucene?

Hi, How to know the details about the application/user who put the SQL Server database in single user mode? Is there any way to identify the user?

@Colin'tHart it doesn't seem like it. More like the previous responses (that they would release the trademarks if some conditions were met). This response seems to say they are releasing it anyway, without expecting anything back.

To me it looks like some power game between the various orgs (PostgresEU, FundationPostgres, PostgresCA)

@Vérace maybe your friend Michael is involved as well ;)

Michael Stonebraker?

I admire the man a lot - he laid the foundations in the 70-s for a db that's still a leader...

yeah

looks like they co-authored some blog post last year

enterprisedb.com/blog/…

Got it...thank you and sorry about it...

@Vérace it's certainly not off topic to ask that question here ;)

but it's not sure you'll get an answer immediately @MuraliDharDarshan Asking on the main site is a safer bet, agree with Verace there

ok, just did:

Thank you guys :)

@MuraliDharDarshan I would hope that you can get the info from audit logs docs.microsoft.com/en-us/sql/relational-databases/security/…

but I am not at all sure (if that is always enabled, if ALTER DATABASE is logged there, etc).

Hmm... this by Álvaro Hernández appears more interesting than I first thought. His StackGres (err..., um...) stack is Open Source under the AGPL - a licence I admire exactly because it stops the big boys (and we know how they are) essentially free-riding on the efforts of the Open Source community! Some OS companies have changed to the AGPL and/or other Source Available But Not For Free (sensu Stallman)... licences. Maybe there's been a falling out between Ongen and AWS?

As the self-appointed official PostgreSQL Ayatollahᵀᴹ, I'm hereby issuing a fatwa against AWS, Jeff Bezos, the Fundacion PostgreSQL and anyone else blaspheming against the sacred name of our server and its prophets!

oh boy, all the 3 and 4-letter agencies will start following the channel now

@MuraliDharDarshan Not unless some form of auditing was eneabled.

Well you just have to write about wanting to disassemble stuff that shouldn't be disassembled, with the aid of chemistry and talking about how you admire some religion or not, then add a pinch of extremism and you're well on your way to being tracked by some distinguished agency.

However, if you were to write that you loved red meat, then the vegans will have you first for brekkers.

@JohnK.N. or, indeed, employing exothermic reactions in the process of disassembling read meat.

Morning heap

@ConcernedOfTunbridgeWells Strangely, just one is an older AS400 based solution. Two DB2/ZOS systems that are more right than wrong, but we still somehow have two of them. Most stuff gets written into a horrible EAV/MongoDB hybrid that makes me question how anything works.

@bbaird I worked at a personal lines site that had separate policy admin systems for home, motor, breakdown and either pet or travel cover. One ran off an IBM mainframe, a couple had Oracle in the back end and I can't remember what the last one was.

Life and pensions systems are quite different as well, and personal and commercial and speciality systems tend to be quite different critters.

@ConcernedOfTunbridgeWells "We need to offer this new product - let's update the data model so we can start offering it" "Whoa... we can't... uhhh... downtime? Downtime." "Guess we better build a new system!"

"EAV/MongoDB hybrid" sounds like a 2-headed dog in the entrance of hell ;)

Five years later: "Why is analytics so expensive? Why is our data warehouse a disaster?! Well, better go build a data lake!"

@bbaird I don't see a lot of homebrew policy admin systems these days. Most are COTS or hacked about COTS products.

@bbaird the site I'm working at right now did just this, although apparently the data warehouse is at T+7 years.

@ConcernedOfTunbridgeWells There really need to be better offerings around this because everything I've seen built is terrible in its own way and horrendously bad in terms of data quality.

@bbaird Data quality is mostly issues with process. Unfortunately getting underwriters to enter stuff that's fit for purpose can be hit and miss.

I did a bunch of work on an application called DQPro that's designed to put a workflow and tracking mechanism about this.

@ConcernedOfTunbridgeWells It's 50/50, few get out of sequence endorsements correct or allow invalid transactions to be committed (endorsement on a cancelled policy, reissue of an active policy, etc.)

With brokers it's even worse. One site I worked at had a system hanging off a messaging hub where they had turned all the validation off. If your system rejected a quote or a renewal the brokers would go to another carrier.

It's a "do business at all costs" mentality

@bbaird Then interpreting the resulting mess like Cancellation-Endorsement-Cancellation-Endorsement-Endorsement-Endorsement as someone frigs about with the system trying to enter the cancellation fee correctly.

It's less problematic on retail products because individual items of crap are for small amounts. On bigger commercial risks it's more of an issue.

It's a big issue around compliance, at least in the US.

Like, I wasted 6 months of my life because no one though that a rating element should have a foreign key constraint to the rating model used. Surprise, surprise, now you've got half a department calculating refunds, issuing data fixes, and you STILL have the risk.

There are compliance implications here but the regulators aren't as bad as there, I think. Unless you're working for a US owned company and have to deal with Sarbox.

@ConcernedOfTunbridgeWells SOX is a concern mostly on the ledger side of the house, but it's mostly the 51 regulatory agencies (states + district of columbia) that cause the most headaches.

@bbaird My current gig are about to implement a pricing platform.

How frightened are you?

@bbaird I've pulled consulting firms out of the shit before.

However, with this one I wouldn't be surprised to see the Law of Third Attempts™ come into play.

I don't think I will have enough control over the project to prevent them from screwing things up if they're going to.

But, I will be able to write up the minimum data standards.

I suspect that the more likely income will be me telling folks something to the effect of "I told you this would happen."

@ConcernedOfTunbridgeWells Strange how they never like to hear that

OTOH it might keep me on an outside-IR35 gig for a couple of years.

@ConcernedOfTunbridgeWells interesting typo

assuming "income" was meant to be "outcome"

it does work though

Oh goody, I get a new job title?

If you have a (temp) table with an identity column and you INSERT INTO #table (blah) SELECT blah ORDER BY blah; -- is the identity assignment actually guaranteed to be in that order? I seem to recall that that is not actually guaranteed.

@CadeRoux What exactly do we mean by "order"?

That the identity integer assignments are also in the order of the ORDER BY

Yes, I do not know why they didn't instead do INSERT INTO blah(blah) SELECT Sequence = ROW_NUMBER() OVER (ORDER BY) which would have made a lot more sense.

The later loop (yes) also relies on the number starting at 1 and seems to rely on not having any gaps, although gaps might not cause it to break.

ok so job title update somehow turned into a salary increase? I'm not going to complain

The circus pays well, what can I say?

@CadeRoux Yes it is one of the few guarantees (from docs.microsoft.com/en-us/archive/blogs/sqltips/… by Conor Cunningham, Query Optimizer Development Lead at the time)

> INSERT queries that use SELECT with ORDER BY to populate rows guarantees how identity values are computed but not the order in which the rows are inserted

It isn't your lawn or your hair

@AaronBertrand strikes again 😁

@PaulWhite thx

So to be clear, the ORDER BY determines the order in which rows are assigned an identity value. The normal caveats about those values being contiguous etc. apply. I don't know that I would build a bank based on that being the case forever though.

I've optimized the loop all away, working on the temp table and the string splitting. They were using PARSENAME to split delimited strings, but as you might expect this fails if the string have [] in them, because PARSENAME expects SQL Server names to really be object names...

@PaulWhite Yeah, I eliminated that with ROW_NUMBER() OVER (ORDER BY) already, so it is guaranteed to be sequential

Cool. The one nice thing about IDENTITY is it cannot be updated.

They had an explicit set identity_insert #tempSessions off in here for some weird reason

Interesting, I got to run it on a customer data and it went from 97 seconds to 1 second. Guess I earned my paycheck this week.

WHAT IS UP, CAPT. CADE

Trash

Trash everywhere

At least someone's happy at the prospect... 😁

'Morning all