@ypercubeᵀᴹ Maybe, but you got a cake (with raisins!) so that's all right then

There is a definite negative correlation between the number of exclamation points in a question and its overall quality.

Also, raisins.

Has anyone else been playing with graph in SQL Server 2017? I'm a bit underwhelmed so far...

@MichaelGreen Pretty basic, is it?

Yeah. Node and edges are just plain old tables with additional system-generated surrogate columns. Much like you'd have built it yourself any time since 1985. Join syntax is old-school from A, B where A.c1 = b.c1, except using the fancy "new" MATCH syntax familiar to anyone who's touched Neo4j. I've haven't been able to get multi-edge queries of indeterminate length to work, and am not getting any response from MS on how it's supposed to.

I get that it's V1 software (or v0.1) and I can see how they could build to an impressive, integrated product. When a single query can span in-memory, column store, row store, graph, xml and JSON it will be a great achievement (or a dog's breakfast, depending on your point of view).

Currently all the work in is user-space, though. To do real "graph" work one would have to learn all the techniques and algorithms and implement them anew in each application. When that's been pushed into the query processor it will be great. At the moment it feels as primitive as the DB only returning a list of tuples and it's up to the application to join and filter the list. I'm looking forward to see what makes it into CTP 3 and SSvr 2019's going to be awesome!

Thanks!

@War Morning, putting binaries isn't a issue

I am not sure how SP going to pick up Web API, with WCF it requires .svc file

good morning, midday, afternoon, evening, night wherever you are

evening

@PaulWhite evening of 26th ?

yes

you are already in future ? lol as it's 26th morning here ;d

Morning

love..love..love..

Howdy folks :)

hello

@Mathematics having found this on your own, you now know :) channel9.msdn.com/Events/SharePoint-Conference/2014/SPC404

@avitex Nice but this is not a quickie question! It's a common issue, with different solutions. See the subtypes tag.

@ypercubeᵀᴹ Thanks! I didn't know the term for what I was searching for :)

supertype/ subtype is one way of referring to this kind of design problems

See this answer as well:

@ypercubeᵀᴹ Interesting. Thanks.

@ypercubeᵀᴹ That's a great resource. Basically I have in this case 3 sub-types, and not many fields so I might just split them out

Shocking network latency if they were really asking about 2016 CTP 2 ;)

@PaulWhite yeah, my comment should be about 2017 CTP, not 2016.

:36964744 Yes.

And dbfiddle is running it.

A ping time of ~2 years would be poor even by New Zealand standards.

@PaulWhite thnx. I noticed a minute after asking ;)

No worries.

@PaulWhite do you have a version of SQL Server 2017 running?

I had a weird issue when using SQL Server vNext when preparing my session for SQL Saturday

Rant of the day: Master Data Services is ever so cheap and nasty.

@Lamak No not yet. I'm resisting VMs and not quite comfortable with 2017 SxS.

darn it

@Lamak What was the issue?

@PaulWhite a very small issue. I was trying to try the new TRANSLATE function, but it said it wasn't a recognized function

didn't work

@Lamak And it wasn't a compatibility level thing?

@PaulWhite I mean...I checked that

and every other new function worked

it was only translate that didn't

Hm that is odd then. Can you still repro? Reported it?

I had to use dbfiddle to show that part (because translate worked there)

@PaulWhite I already uninstalled it

it was kind of a hit though, because people loved dbfiddle lol

Yes I now remember you mentioning that to Jack.

@Lamak does it requires any language extension?

@McNets it's not that kind of translation :)

@PaulWhite yup, I did

@Lamak ok

@McNets docs.microsoft.com/en-us/sql/t-sql/functions/…

It can be very useful

@Lamak Do you still have the code you were running, and the error message you received?

It's similar to a replace() function

@PaulWhite no, sorry....but just using the code on the docs wouldn't work

the message was the typical: 'translate' is not a recognized function name.

@AndyK it is like an insane query

@McNets it's like replace on steroids

status-norepro

Got there in the end :)

lol @McNets

Community driven Enhancements in SQL Server 2017

Select into on filegroup is a nice one.

Bonus points if you could do select into onto a partition scheme.

Making master data services not suck would also be nice but that's probably a bit too much to ask.

There are tons of things I would have liked to have seen as well, but hey I'll take it.

There are a few nice things in 2017 but mostly v1/intent things.

2017 is mostly about Linux I guess.

And fair enough for that. Making it work slightly faster on a new OS is quite an achievement.

Maybe the should hire a couple of kernel hackers and see if they can get the kernel tweaked to help with their performance. That would be a major flag-waving exercise and show of good faith in Linux circles.

How do we know they didn't already do that? ;)

@PaulWhite Last year, when I opened the box in which the Ubuntu 16.04 LTS release arrived, small Bill Gateses jumped out of it

Well there you go. Proof!

More seriously, if this releasing new things at alpha/early beta and hoping for feedback to improve the next iteration thing is going to become permanent, well it might take some getting used to.

Not the *nix stuff, for clarity.

For example the graph stuff Michael Green was talking about earlier.

@TomV Yep! Will do so, thanks. I appreciate it.

should we give this one more chance? I might be persuaded, @PaulWhite

btw, I did respond to your question answer on meta

@MaxVernon it has my vote already.

I'm tending to agree with re-open only because I added an answer as well as the initial VtC. I still think the question sucks.

if you want to insert an item into the middle of a linked list, you MUST update the prior item and the next item so they point to the new item, instead of each other. There is no magic involved in that, unless I'm just saying "get off my lawn" without realizing it.

I'm fuzzy on what they're looking for. Today the order is [C, B, D, A] and tomorrow I might want it as [A, B, D, C]. Beyond explicitly asking for that, I have no idea how a set based concept like a database is going to solve their problem

@MaxVernon Thank you!

no worries mate. I've been very busy the past few days.

I'll try harder in future not to answer crap questions like that.

@MaxVernon did you post an answer in SO?, leave that to us plebeians

I mostly disagree with the reason to VtC at this point, shopping list would be a different matter

But it's not my domain, so I voted to reopen to give the community a second shot, if anybody closes it as off-topic again instead of unclear I'm fine with that

I suppose there's a way to reword it though. What types of databases instead of what database could make a difference

if you get rid of a lot of the fluff it seems very similar to this question? dba.stackexchange.com/questions/148732/…

I think his problem is that he doesn't realize that he can create APIs to do what he wants

and as long as you have an API that performs well, the "clunky" bits hidden underneath it don't really matter

@JoeObbish good point

I left a comment on the question

@JackDouglas dbfiddle.uk seems to not respond occasionally

@ypercubeᵀᴹ @JackDouglas yep

we need young @JackDouglas to work more on it ;)

should this edit be accepted? dba.stackexchange.com/questions/172081/…

And here I thought I had broken @JackDouglas fine fiddle with my oh-so-clever query

In fact it clarify a bit the question.

350 M rows into a GROUPING SETS operator is a worse time than I thought...

it should be fine now, hoping it won't break again tomorrow

I found my first bug in the Linux kernel :)

first? aren't we cocky!