The Heap™ – Consultancy ©®

ypercube

06:01

hi i m using sql server 2008 db.index will not allow duplication in the index but we need to avoid duplication in table itself — anubhooti 17 hours ago

bangs head in wall ...

Mark Sinkinson

08:16

@JamesLupolt There's one or two that go to them. But it's generally discouraged. And they don't actively disclose they're recruiters

dezso

08:28

0

Q: How to Pass table type as Parameter in T-Sql like Postgresql

In my Company like to move pl/Pgsql to T-Sql. In pgsql function we are create and pass type as parameter. Same option had in T-Sql or not. If not present it have any alternative to same like pgsql. Postgesql Function CREATE OR REPLACE FUNCTION testfn(test) RETURNS text AS $$ DECLARE xsal...

Am I the only one who does not understand what this is about?

Paul White

@dezso He's just asking about TVPs, right?

Converting Postgres code to SQL Server.

dezso

@PaulWhite now this is never mentioned

(tpostgres.org/se to increase the confusion)

but if it is so, I think it's TVP, right

or TVF, dunno

Paul White

@dezso That's how I read it. I might be wrong.

dezso

@PaulWhite I've asked it

BTW, is the code in the answer syntactically correct?

Paul White

@a_horse_with_no_name sir check my updation — SATSON 1 hour ago

ypercube

08:36

That's how I read it, too.

> i need like same as in t-sql functions

dezso

@Phil apparently, I've posted incorrect code in my comment - now fixed it in your answer, please check

Paul White

09:26

@dezso No.

ConcernedOfTunbridgeWells

Morning all

Paul White

Morning!

Mark Sinkinson

09:43

Morning

@ConcernedOfTunbridgeWells I could have sworn I saw your twin in an office on Bishopsgate the other day

ConcernedOfTunbridgeWells

09:59

None of that heretical escape-meta-alt-ctrl-shift bollocks ...

ConcernedOfTunbridgeWells

10:19

@MarkSinkinson Ah, my evil twin Skippy.

5

ConcernedOfTunbridgeWells

13:20

I just prodded @MarkStoreySmith about dates for drinkies and he says he's back 22/23 october and 26/27 November. @JamesLupolt @JackDouglas @MarkSinkinson @ypercube @MartinSmith @Phil - Is Fri 23 Oct viable for drinkies?

4

billinkc

No invite for me? Fine, just for that I'm headed to Portland (OR) to be even further away from such shenanigans

ConcernedOfTunbridgeWells

@billinkc You're most welcome to come across the ditch. No Concorde these days but still ...

billinkc

@ConcernedOfTunbridgeWells No, no. Too late now. I wouldn't want to be a burden

Marian

13:40

@billinkc You are anyway, but why'd you care? :)

3

billinkc

:'(

ConcernedOfTunbridgeWells

@billinkc :p

@billinkc You're never a burden. Why not come to sunny London, home of the liquid lunch, lots of queen-themed touristy crap and a big ferris wheel?

Plus, several buildings with clever names like 'Walkie Talkie', 'Gherkin' and 'Cheese Grater', and my personal favourite: 'The Dick'.

Note the asymmetry. Michaelangelo would have been proud.

Phil

13:57

@ConcernedOfTunbridgeWells 26/27 Nov for me. Can't do 22/23 Oct

Call it an early Heap Christmas :-)

Phil

14:12

Is there an easy way of calling a web service from T-SQL, or should i just forget about that & do it elsewhere :-)

2008r2, I'm afraid

JNK

@ConcernedOfTunbridgeWells I would love to come to London

Aaron Bertrand

14:27

@Phil do it elsewhere. There are potential hacks, but why? How is the T-SQL ultimately getting called?

Phil

@AaronBertrand Got a BPM product sitting on the server that can execute things based on app triggers. It's be easier to do it in the DB than go to an external prog

Aaron Bertrand

@Phil hitting a web service isn't something you should be doing from inside SQL Server, and the ways to do it are risky. E.g. enable xp_cmdshell and hit some program to do it anyway, or use sp_OA, e.g. sp_OACreate 'MSXML2.XMLHTTP'

xp_cmdshell opens up plenty of security issues, and sp_OA is known to have memory leaks.

You can write a C# app in 5 minutes that polls a queue table and hits the web service URLs that BPM can put into the queue table

Also see this, but ugh. The above sounds much cleaner to me.

Phil

Cheers, I'll go down that route. I wrote a P.O.C. perl script first to test the API out, but time to do it properly now :)

Aaron Bertrand

Perl made me shudder

Phil

Heh, it's a hard thing to give up. I'm not used to this Windows world yet

ConcernedOfTunbridgeWells

14:44

@JNK There are direct Boston-London flights if you feel that way inclined. Might be less sodding about than going via New York.

Option 1: Do a .net sproc. Cons are you've got to deploy it and debug it server side
Option 2: Do something client-side - Powershell or anything .net-ish could do this. Build a log table for requests in the DB and have something client-side poll it.

In a pinch you could host the poller in an SSIS package via a script task, but this also puts you in the position of having to debug the thing in situ within the package.

@AaronBertrand It has this effect on a lot of people, although it's bark is worse than its bite. Debugging legacy perl code, however, is a completely different story.

Phil

@ConcernedOfTunbridgeWells Looks easy in Powershell. Just googling to find out how the hell to do Basic HTTP Authentication

swasheck

@Phil piece of crap?

2

Phil

@swasheck Yeah, that and Proof Of Concept. Same shit, you're right :)

swasheck

@Phil right. most of my POCs are PsOC

Mark Sinkinson

SSIS is not bad with web services if you want to go down that route

@ConcernedOfTunbridgeWells Can't do next week, so November is better for me

Phil

14:56

Awesome, Powershell script done. The BPM thingy can launch it, sorted

Thanks for the advice

JNK

@ConcernedOfTunbridgeWells Someday! problem is bringing the whole fam etc etc

jcolebrand

@Erik you're in the top 80 of 3k people ...

I would say you're pretty senior.

Hell, at least one of our moderators is not in the top 20 of our users :p

James Lupolt

15:54

@ConcernedOfTunbridgeWells 23 Oct I'll be at PASS Summit in Seattle but don't let my absence stop you. The Nov dates should be fine for me.

swasheck

@JamesLupolt i'm not arriving until the 26th

James Lupolt

@swasheck You aren't going to any precons I guess?

swasheck

@JamesLupolt no. i'm not. are you?

James Lupolt

yes

tripp's and delaney's

swasheck

which?

James Lupolt

16:02

they seemed like the ones where i'd get the highest ratio of stuff i don't know yet

swasheck

is msdn dragging for others?

James Lupolt

not me

swasheck

interesting.

@JamesLupolt my coworker will be in tripp's. he's leaving early in the week so he got approved for a precon

James Lupolt

16:34

Cool

Erik

7

A: canned responses for "leave a comment"

I was the one who flagged your answer1. Acknowledging the fact that you're breaking the rules doesn't mean you get to break the rules. The flag merely agreed with your assessment that you were breaking the rules. The flag isn't meant to punish you. It is meant to be helpful. By flagging the ...

James Lupolt

One of the developers? Or another DBA?

Erik

@jcolebrand Do you mean my comment here?

jcolebrand

yes

Erik

@jcolebrand lol I still think most if not all of you guys are better BDA's than me. The fact that I'm 121 overall as far as rep goes just means you'll let any a-hole in here. :)

jcolebrand

16:37

@Erik lol, I mean, my dba skills are midfield

my moderator skills however

also mid-field

HA!

Erik

lol

Where did you get the 80 out of 3k number?

jcolebrand

I just wanted to help make this site better for dbas like me when we first started, and I think I helped do that, and now I play elder statesman and still show up every so often

@Erik I ranked all users by all time by rep, and paged through till I found your name. You gotta click a few buttons to make that happen. If I can math, the users page has 20 users on it and you're on the 4th page

I probably can't math tho, I've been strung out on converting C# to SQL for the past few days

swasheck

@JamesLupolt another dba. smart guy. his name is brandon.

Erik

@jcolebrand I think that is why I show up so well on the leader boards. Lots of really smart people take more of a "elder statesman" role.

@jcolebrand At least you aren't being required to translate the other direction. That would be much worse.

jcolebrand

@Erik right

Erik

16:43

@jcolebrand Related to this I was surprised to see how few of the people with a precog badge have serious rep and/or still participate regularly. dba.stackexchange.com/help/badges/61/precognitive

Same goes for the beta group, with you being one of the notable exceptions. dba.stackexchange.com/help/badges/16/beta

jcolebrand

Yeah, been a little while since then lol

Erik

@jcolebrand I for one am grateful for your efforts.

Andriy M

@Erik What surprises me is that the badge was still awarded in 2014. Shouldn't the awarding have been done once and for all in 2011?

Erik

@AndriyM That is a real head scratcher.... Maybe they followed before it entered beta, but first visited the site after 2011? Who knows...

swasheck

16:59

NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO‌OOOOOOOO

Andriy M

@Erik Or maybe once a site leaves Area 51, it gets its own user base (as opposed to the Area 51 user base) and people have to register specifically for the new site, and maybe not everyone did so for DBA.SE but the system recognised them when they did later

swasheck

listening to a great song and then i clicked refresh instead of the tab

Andriy M

@Erik Hmm, I guess I just reiterated the same you said :)

Erik

@AndriyM Agreed but you stated it better than I did. :)

srutzky

17:16

@Phil Yes, there is a fairly easy way. I believe the following answer on S.O. addresses this:

3

A: Call web service from SQL CLR?

Please do not use the sp_OA* OLE Automation procedures. They do not appear to be officially deprecated, but SQLCLR replaces both the OLE Automation procedures as well as Extended Stored Procedures. Yes, this can be done easily enough in SQLCLR. You can find examples on using WCF (as shown in @Co...

As I mention in that answer, you are probably better off creating and/or using a generic Web Request/Response and handling the creation and parsing of the XML yourself. This way you can reuse the same proc for any number of calls and won't ever need to change it when the service changes.

As I don't mention in that answer, of the other one I linked to in that answer, if you don't want to bother coding this yourself, a SQLCLR stored procedure is available in the SQL# library (which I am the author of), but it is only in the Full version, and not available in the Free version. But at least you have some options :)

billinkc

18:33

Really, a copy tag? dba.stackexchange.com/questions/tagged/copy To the burnination mobile

swasheck

@billinkc postgresql.org/docs/9.2/static/sql-copy.html

billinkc

But if you look at the current usage, it's shite

swasheck

that's like saying we shouldnt use sql because our developers do horrible things with it

billinkc

0

Q: PostgreSQL: How does the implicit index work when using "COPY" for loading bulk rows into a newly created table?

What I know for fast loading bulk rows into a new table is: Create the table(without creating the index) Use "COPY" to load data from a file into the table Create the index What if I added a primary key when creating the table, according to PostgreSQL: "Adding a primary key will automatically...

postgresql index constraint copy

That appears to be the only question that actually talks about the Copy capability. The other questions appear to cover How do I copy the data/table/database from A to B

Tom V

19:09

@Erik I'm not sure about the a-hole thing, it's mostly a clown-pit I guess

I haven't seen many jerks around

at least that's what I consider myself, more of a clown than knowledgeable

Erik

@TomV I agree I don't see any real jerks.

Tom V

except @billinkc

/JK

billinkc

@TomV No need to add that qualifier, I accept my jerkiness

Erik

@TomV my a-hole comment was more of a colorful way of saying anyone, and probably should have been substituted for "joker"

Tom V

@billinkc that WAS the qualifier /JerKiness

Erik

19:15

@TomV I enjoyed your "clown-pit" comment by the way. I hadn't heard that before. For fun I looked it up but the top hits were all for a band called Slipknot, and I assume they were referencing obliquely towards that band's feud/rivalry with another band called Insane Clown Posse

billinkc

/slowclap

Tom V

rimshot.wav

@Erik I didn't even really know anything about that band except hearing it existed and was big with a special kind of crowd

Andriy M

@billinkc My English dyslexia first caused me to read that as "slowslap"

billinkc

I'd use a much more vigorous slap given the opportunity

srutzky

@Erik and @TomV Not everyday I see Slipknot mentioned ;-)

Erik

19:22

dba.stackexchange.com/review/suggested-edits/57566 I rejected this because I didn't see confirmation in a comment. Since @MaxVernon approved it then I might have missed a deleted comment and/or something else

Tom V

@billinkc I'm not into that kinky stuff you know

Erik

@srutzky lol we're an eclectic group here at the Heap

billinkc

Pity

Tom V

LOL

Erik

@TomV Not really my type of music either, but I'm one of the weirdos that enjoys classical music. No accounting for taste I guess.

James Lupolt

19:25

@swasheck OK cool. Maybe we'll run into each other. I think there will be like 100-300 people in the precon.

Tom V

@Erik Ha I was raised with that music, kind of let it go when I was in my late teens but still appreciate it, especially in the car for long drives

Erik

chat.stackexchange.com/transcript/message/24708549#24708549 The user mentioned the link in a cross posted question that @MaxVernon found here stackoverflow.com/q/33125492/5190842

srutzky

@Erik nothing really wrong with most any type of music...liking classical doesn't make you wierd..

Tom V

especially "modern" classic music, like bela bartok and his contemporaries

Erik

@srutzky that's true. I find some genre's of music hard to listen to due to the content of their lyrics, but overall I agree with you.

Tom V

19:31

I spent a lot of time refurbishing homes with my grandfather who played in semi-famous orchestras (in belgium) when I was young and while working all I heard where violins and pauks and trumpets :)

Erik

@TomV I'll have to check him out. I might of heard some of his music without knowing it. Mostly I listen without knowing who is playing.

@TomV That is very cool

swasheck

adding (2) indexes now leads to deadlocking ... make sense to anyone?

Erik

@srutzky And how often do you hear Slipknot mentioned right before a discussion of classical music? :)

Tom V

@Erik I'm not sure how to translate it, but they had room ensembles, trios and quartets rehearsing in his home while I stayed there

Erik

@swasheck That is related what @MaxVernon was talking about a couple of days ago. A new index can lead to a new query plan which can ripple consequences through the system

swasheck

19:36

@Erik while true, i'm not sure that i've ever encountered a string of deadlocks being one of these ripples

Tom V

@swasheck that's my (although limited) experience too, in theory adding an index has an overhead and a risk, but most of the time it helps the search pattern the app/your users have and doesn't lead to disastrous problems as long as you don't exaggerate

I did see a parallelism/deadlock issue not too long ago after adding an index on a fairly large table in a fairly specific case

swasheck

@TomV on write?

Max Vernon

@swasheck I'd certainly be interested to see how/why that is the outcome.

Tom V

@swasheck yes, on update

swasheck

parallelism makes some sense ... if two write processes have gone parallel and one has a lock on page a while the other has a lock on page b and their parallel threads need access to each other's locked page. or something.

Tom V

19:45

that was the graph at the time

I never fully grasped it, but that is my ignorance I guess

swasheck

that's a neat one

Tom V

@swasheck I never fully understood the problem but it's a central history table written to by multiple threads and multiple servers so I wouldn't mind grasping the real problem actually, maybe that should be a main site question

swasheck

@TomV indeed. i'd upvote it.

Erik

@swasheck me too

Tom V

I'll try to create a fully documented question tomorrow or friday but it's kind of difficult to have a tested solution as it's an old problem, so that might get in the way of accepting responses or testing suggestions

Erik

20:00

@TomV Or maybe @swasheck could write the question since he seems to be experiencing the same thing

Tom V

it was a batch job history table and that history has been cleaned up by now and maxDOP has been set to 1 as per the microsoft dynamics ax recommendations (yuk)

@Erik possibly better yes, I'll upvote that but I don't see how my graph could be posted as an answer rather than a comment

Erik

@TomV True but if we sweet talk @swasheck enough he might create his own pretty graph.

@TomV I'm impressed that you were able to pull out that graph so quickly for an old problem by the way

swasheck

unfortunately i'm not allowed to do so without some serious anonymization

Tom V

@Erik I pulled it from a report from back in that day

I took a screenshot back then

@swasheck could use a trial plan explorer or talk @AaronBertrand into giving him a license so he can create his own

Erik

@swasheck joke --> That is why you should always use the naming pattern: col1, col2, table1, table2, etc .... So you don't have to worry about anonymizing just in case you find a weird bug

swasheck

20:06

@TomV we have it

Tom V

@swasheck then you should be able to (depending on version) create a graph like that based on the default trace output

swasheck

@TomV except it's on a qa server that's not being monitored by plan advisor :)

but yeah ...

Tom V

that shouldn't matter

sec, let me google, I didn't find out myself ;)

swasheck

you can anonymize a deadlock graph with plan explorer?

Tom V

no idea but this article helped me capture and graph it simple-talk.com/sql/database-administration/…

but mine isn't anonymous as you can see the hostnames and application names

but @AaronBertrand would be the go to guy for anything related to sqlsentry stuff

Aaron Bertrand

20:16

@swasheck No, we've only implemented that for query plans

swasheck

@AaronBertrand i didnt think so :)

@TomV right. we're familiar with his level of support :)

Tom V

@AaronBertrand that shouldn't be too hard of a feature request to implement if enough customers want it i guess

swasheck

not sure how useful an anonymous deadlock graph would be :)

Tom V

@swasheck who would care if it said table1 and table2

Aaron Bertrand

@TomV well it could get confusing if the deadlock is single-table. I don't know that anybody has ever asked for that functionality, or if we would even take it seriously if they did. :-)

Tom V

20:23

I don't mind putting my graph online but I wouldn't be able to test any solutions as the table has been reduced compared to back then

Aaron Bertrand

And I'm not being flippant, just honest. We created a site specifically for people to share their query plans, and so allowing some kind of anonymization was natural. For deadlock graphs, it wouldn't be extremely difficult to do search/replace in the XML yourself, if you had a scenario where you needed to share your deadlock graph online without revealing table names (which I don't even think is necessary in 99% of the query plan cases, either).

swasheck

@AaronBertrand right. from my perspective it's silly. from my employer's perspective it's a breach of disclosure agreements ... baseless as it is.

Aaron Bertrand

Yep, silly. ZOMG THEY KNOW OUR CUSTOMERS TABLE IS CALLED dbo.Customers!

swasheck

unfortunately, this is a vendor product that we're trying to help limp along which further muddies the whole disclosures thing

Tom V

@AaronBertrand yea I agree but a signed NDA is a signed NDA I understand too

swasheck

20:28

apparently somebody signed something dumb

Aaron Bertrand

@TomV Yeah I get it but it shouldn't be written so broadly.

swasheck

well i'm stuck here until someone like microsoft or sqlsentry buys out my profit sharing and vesting schedule. and i'm not worth that

Aaron Bertrand

Also why Microsoft forces everything to go through Windows Live Login but the MVP summit site uses something completely different is annoying. More annoying: no password reset. Having trouble? is a mailto: link.

swasheck

@AaronBertrand security through hilarity

James Lupolt

@TomV It looks like a bird

Erik

20:32

@swasheck or until you win the lottery....

Tom V

@JamesLupolt are you referring to the image or am I missing a foreign language joke

James Lupolt

The image

Tom V

@JamesLupolt the deadlock chart looks like misery to me

cleaning up the 3 years of history solved most of it

that and not logging succesful job completion

James Lupolt

I've heard Dynamics AX is a lot of fun. Sounds like it's true

Tom V

@JamesLupolt depends I guess, 2012 version without SP is a mess yes

previous versions were ok, later versions also

Aaron Bertrand

20:42

stumbleupon.com/su/2Nsht0/www.collegehumor.com/picture/6687821/…

James Lupolt

Good to know. I suppose you also have to maintain perspective: it isn't SAP.

Tom V

@JamesLupolt I work with it day to day, it's manageable if you don't run into their first attempt at table inheritance in their RTM version i suppose

swasheck

> URL: www.collegehumor.com/picture/6687821/man-with-bizarre-name-arrested

Category: questionable

Tom V

@JamesLupolt any mature ERP is a mess, I have been doing BI on SAP and that was worse

Aaron Bertrand

You people and your paranoid control freak IT departments

Tom V

20:47

but it beats custom dev in many cases

Aaron Bertrand

I'm torn, some of these "do everything" products save time up front, but in the long run...

swasheck

@AaronBertrand it's the nature of the beast, i suppose

Aaron Bertrand

Just like EF and other ORMs and frameworks, sure you get to market faster, but what you bring to market faster is not all that good.

Unless timing is the most important factor, I would rather spend the time and money on developing exactly what we need, having complete control, instead of saving some development time and giving up control / stability / complete feature gaps

Tom V

@AaronBertrand that's not the same, using something which is ok for ledger/inventory/projects/etc out of the box has some value, even if it has its problems

James Lupolt

@TomV The thing about SAP is that it seems like you need to do lots of custom dev to get it to do what you want. At least that's what I'm observing at my current job.

And the custom dev seems to be done by an army of expensive consultants.

Tom V

20:50

@JamesLupolt but think about the dev work if you start from zero

Aaron Bertrand

For smaller pieces, sure, you can plug in things to avoid reinventing the wheel. But for over-arching features and workflows, nobody can build a generic product that does exactly what you need.

@TomV I think it's debatable. Sometimes it takes a week or two to get some of the more advanced suites up and running (think Sharepoint, Dynamics, etc). It's not like you turn them on and don't have to hire developers or system folks to get it running and keep it running.

Tom V

thing is, salespeople try to shoehorn their solution into any business requirement, sometimes the customisations trump the basic solutions, sometimes they don't

@AaronBertrand exactly the point

but for a lot of manufacturing clients the requirements overlap

Aaron Bertrand

@TomV ? You are advocating plugging something in because up-front dev work is too much. I'm suggesting that up-front dev work is probably more of a valuable trade-off than you are giving it credit.

Tom V

@AaronBertrand no but if your overlap is large enough there is no point in redeveloping ledger/mfg/projects

Aaron Bertrand

@TomV So then you already know what the tables look like, and the interfaces to them. Database is done.

Tom V

20:54

it has a tradeoff off course

Aaron Bertrand

Of course. I'm merely suggesting that you don't discount up-front dev work as pointless, because in a lot of cases it is more than worth it.

Tom V

I would say, roughly 60% of our target audience is ok with a standard solution with some customizations

Aaron Bertrand

I was very close to re-writing WordPress without any MySQL / PHP poison in it a few years ago. There still isn't a half-decent blogging platform that works with SQL Server without breaking your back or wanting to kill any PHP script kiddy you see

@TomV Well I think I have higher standards than 60% of your target audience.

Tom V

SMB stuff is mostly uniform in many ways

Aaron Bertrand

@TomV That isn't a strong argument for plugging something in. If anything, it's a good argument for up-front dev work is actually quite easy, because you should know the table design and interfaces with your eyes closed.

Tom V

20:57

even though the solutions provided by ERP partners is still extremely expensive I agree

Aaron Bertrand

So you can build the tables and the customizations up front, instead of getting the thing up and running out of the box, and then figuring out how to customize it (when that's even possible).

You should see what I have to go through when I want something in Sharepoint to behave slightly differently. Not even worth it.

Tom V

sharepoint I agree with

rather, the issues with it

Aaron Bertrand

Yeah I am not advocating writing your own Sharepoint. Just using it as an example of turnkey solutions that aren't easy to work with.

I suspect most of the things you are talking about are far less complex than Sharepoint.

Tom V

but production scheduling is fairly complex and if you can resort to a parametrised semi-standard implementation I'm not necessarely against an established mrp system

@AaronBertrand sharepoint would be simple in ERP terms, I respect your experience and knowledge, but an ERP is a different beast

swasheck

at least i got to mess around with this

select
	qp.session_id,
	qp.request_id,
	qp.thread_id,
	qp.node_id,
	qp.physical_operator_name,
	o.name,
	qp.row_count,
	qp.rewind_count,
	qp.rebind_count,
	qp.end_of_scan_count,
	qp.estimate_row_count,
	qp.elapsed_time_ms,
	qp.cpu_time_ms,
	qp.scan_count,
	qp.logical_read_count,
	qp.physical_read_count,
	wt.wait_duration_ms,
	wt.resource_address,
	wt.resource_description
from sys.dm_exec_query_profiles qp
left join sys.objects o
	on qp.object_id = o.object_id
left join sys.dm_os_waiting_tasks wt

Tom V

21:02

(not wrt databases, but project based manufacturing, build to order, etc)

Aaron Bertrand

@TomV Right but what I'm suggesting is that you don't need a single box that "does all the ERP things" - ERP is a huge mash of largely independent features and entities.

swasheck

but i can't figure out how to do anything with the stupid "timestamp in bigint" thing. thanks, ms

Aaron Bertrand

They're all related to the business but they are modules that can be developed independently for the most part.

Tom V

@AaronBertrand yeah but for an SMB with 200K budget an in house developed solution might be far fetched and sales people sweet talk them into a semi-standard solution with expensive licenses

and I don't necessarily think a custom solution would be cheaper actually

Aaron Bertrand

I don't think you can generalize based on those very specific parameters.

Tom V

21:06

and I will give you the arguement of 200K is a lot of money, and I will give you the argument of 'we can do a lot for that money'

Aaron Bertrand

Well, you can, but you may not be able to get a lot of people to agree that, yeah, you're right, you should always just buy shelfware that kind of does some of that you need

Tom V

but there is a lot of standard functionality in those applications,and I'm not a salesperson and I'll surely agree it's not a brilliant fit for many of them

for example the customer I was in conctact with today, will see him next week, 40 person joint, his sig went from marketing to ict responsible back to marketing

that's not the customer for a custom solution

the customer I was with today OTOH, he was convinced into a solution based on microsofts solution which deviates too much from their requirements and that's a deep and profound mess

it depends as always

in the end it all ends up being unknowledgeable guys (if thats even a word) relying on vendors, and none of the solutions are right

add more biztalk to integrate the crap isn't working either

anyways, i'm out, too bad @AaronBertrand I didn't meet you in utrecht (you were at the table with the rest of the speakers when I saw you) but it's a complex discussion i guess and it should be had IRL, possibly over a beer or lunch

I didn't say hi because I didn't want to interfere

Aaron Bertrand

21:23

@TomV I don't think there's a right answer for all scenarios. I just knee-jerk to the "never build from scratch" attitude because there are many scenarios where doing exactly that is right.

Tom V

@AaronBertrand build from scratch is the right answer in a lot of cases I agree

don't get me wrong

but most of my/our customers are SMB's with a fairly standard way of doing things and having a customizable solution featuring most of their requirements is ok i guess

hence the success of navision, ax and sap one i suppose

but as you say, I don't subscribe to the 'never' or 'always' either

anyway, i really need to close the laptop lid and if you are ever in .nl or .be again we should meet

Aaron Bertrand

Yeah sorry you didn't come over

Would have been fine, I'm never doing anything that important at a speakers' table

Tom V

I was there, but you were in discussion over lunchtime with the rest of the mvp's and i didn't want to interrupt

so I visited the booths :)

did you get to antwerp in a semi-comfortable manner?

Aaron Bertrand

@TomV I did, Dandy drove, good times

Tom V

ok then, public transport would have been a pain

Andriy M

21:36

@swasheck "do anything" like what?

Tom V

talk to you later, let me know when you are in the lowlands again

swasheck

@AndriyM like get an actual datetime

Andriy M

@swasheck But that's a duration, not a point in time. Turning it into a datetime would hardly make sense.

Or maybe I'm misunderstanding

swasheck

Timestamp when open (in milliseconds).

Andriy M

I mean all those *_ms columns are about duration, no?

swasheck

21:39

no ... open_time, first_row_time, last_row_time, close_time

Andriy M

Is the reference point known? Or are normal time equivalents known to figure out the reference point?

swasheck

@AndriyM dunno. nodoc

> Timestamp when open (in milliseconds).

James Lupolt

Do you get anything sensible if you subtract that number of ms from the current datetime? Or add it to Jan 1 1970?

swasheck

@JamesLupolt nonsensical date. used sys.dm_os_sys_info ... and keep getting overflows

James Lupolt

Oh. I wonder if Paul White or Conor Cunningham knows. #sqlhelp any use?

Andriy M

21:46

If I'm not much mistaken, you can't add a bigint number of units in dateadd, but you could divide it by 1000 first

I don't have a 2014 handy so I've no idea what those numbers look like. Could you share some here?

swasheck

@AndriyM yeah

i'd need to set things up again ... maybe tomorrow. i'm too disappointed in my job right now to even care ... so i'm probably just going to close up shop.

Andriy M

Wait, you mentioned dm_os_sys_info, I'll see what it looks like there...

swasheck

it's a bigint

:D

Aaron Bertrand

SELECT DATEADD(SECOND, column_in_query_profiles/1000.0, sqlserver_start_time)
FROM sys.dm_os_sys_info

swasheck

@AaronBertrand ah. the sql_server_start_time was the missing link. TA!

Aaron Bertrand

21:52

That was only a guess and I've only done marginal confirmation.

swasheck

oh

it's somethign to try tomorrow when i dont hate my job

James Lupolt

22:18

slideshare.net/chris1adkin/…

Transcript for

The Heap™ – Consultancy ©®