« first day (782 days earlier)      last day (4383 days later) » 

00:03
@dezso Oh, dear. Is it web scale?
00:40
Good work, Microsoft! theregister.co.uk/2013/02/22/…
2
 
4 hours later…
 
6 hours later…
10:22
@Phil And there you have it:
Cloud computing: Store your data with the lowest bidder.
2
 
2 hours later…
12:27
@ConcernedOfTunbridgeWells I actually have a customer which uses AWS to run the middle tier but keeps their db servers inhouse
exactly for that reason
 
3 hours later…
15:11
@ChrisTravers Every time you buy an outsourced service there is a middle manager somewhere in the chain trying to minimise the amount of work they do and maximise their margins.
@ConcernedOfTunbridgeWells Of course.
That's one reason I am quite skeptical of dbaas anywhere outside of development
@ChrisTravers dbaas -> database as a service?
 
1 hour later…
16:26
@ypercube It was very silly of me to confuse SELECT with DELETE. Sorry.
No problem
 
2 hours later…
17:58
@ConcernedOfTunbridgeWells are we invited?—I didn't get an email
or can anyone just sign up from that link
@Jack thee page "You're invited" at the top, so I guess the latter.
true, but I wasn't sure the link was public :) I think you are right but COTW said he got an email...
His profile has "London", maybe that's why he got email invitation (just a guess).
 
3 hours later…
20:48
Oh my, can someone explain this behahiour in SQL-Server?
SELECT 4 + ('25' + '6')  AS result ;
gives 260 ...
there is some logic behind that
'25' + '6' = string concat to '256'. 4 + '256' = integer promotion to 260
Does that make sense?
Yes, it does, I undesrtand how it works.
As to why, god only knows
I don't understand why they made it work that way..
(you answered before my question :)
my personal theory is that any type conversion should be explicit not implicit but then I'm old-fashioned like that
21:00
But once you have an operator (+) that does 2 jobs, you have a problem. I guess they made this choice.
well no, they could avoid it by disallowing implicit conversion and requiring an explicit CAST or CONVERT
Anyway, I came to this issue, because I was wondering why '' is converted into 0. Which is also strange.
SELECT CONVERT(int, '') shows 0
21:33
Chances are that was originally a bug back in the day but it's become a feature and is now required for backwards compatibility
22:08
@ypercube Are you asking why the rules of data precedence are the way they are, or why the + operator is overloaded?
In SQL Server 2012, you can use CONCAT:
SELECT 4 + CONCAT('25', '6')  AS result;
Not sure I am answering the right point though.
I'm asking why these implicit conversions happen, basically.
@ypercube And given an explicit conversion from '', what would you expect? NULL? An error? Zero seems quite natural to me, but then I might just be too used to SQL Server :)
@ypercube T-SQL isn't strongly typed.
I was expecting error ( i guess)
Which is basically me saying, "because it does".
T-SQL likes to keep people guessing.
I would much prefer explicit conversions always, but I can see how that would irritate people, and might not always be practical given the desperately hard rules for types derived from say arithmetic on decimals.
For example, who can predict the typed result of DEC(6,3) / DEC(8,2)?
Someone did explain why numerics have a higher precedence than string types once, and I remember finding the explanation convincing, but I'll be buggered if I can remember the logic of it.
OK. Now, another question, is the CONVERT(int, '') to 0 documented? I couldn't find it.
22:20
@ypercube Don't think so. Though it is documented that conversion of an empty string to decimal/numeric throws an error.
SELECT CONVERT(decimal(9), '');
Msg 8114, Level 16, State 5, Line 1
Error converting data type varchar to numeric.
There's no shortage of quirkiness though:
SELECT CONVERT(char(1), 23);
Result: "*" < intuitive?
@PaulWhite Interesting! "Works" with implicit conversion as well: SELECT 1 + '' is fine, SELECT 1.0 + '' throws an error.
@AndriyM Interesting is probably a fair description of a lot of the implicit conversion behaviour.
@SimonRigharts was probably right that most of this is just compatibility with questionable decisions made in (much) earlier releases.
Can you imagine testing this stuff for regressions?
@PaulWhite I'd have a hard time keeping the intended behaviours straight, personally
@PaulWhite Also, I think I asked you this once before - do you go to the Wellington SSUG meetings?
22:38
@SimonRigharts Yes you did, and no I don't, but I probably should.
@JackDouglas I got an email invite from somebody called Bethany Marzewski. She said they only have space for 75 people or so so 'please don't tweet.' However, I doubt that anybody would get their nose out of joint if you turned up. If you do want to come you should probably RSVP on the form though.
@jcolebrand Do you know if the London shindig is invite-only or can anybody RSVP and turn up?
for SE?
@jcolebrand Yes.
I got an invite, but I don't think @JackDouglas did.
It should be anybody can RSVP on account of it's on the site at HTTP://StackExchange.com/London
@jcolebrand The email did say they don't have much room. However, I doubt anybody is going to mind Jack and the DBA.SE regulars turning up.
23:21
Nobody holding an SE business card is going to get turned away, I expect. ;-)
@jcolebrand That would be my guess as well - unless they get a great horde of people turning up and gatecrashing the place.
Do you have any idea who Bethany Marzewski is?
@ypercube Maybe.
The profile is consistent with somebody who might send out invites to the shindig.
@PaulWhite Seems to be more about the hobnobbing than the actual presentations, but maybe that's just me
23:30
"Bethany Marzewski author"
@ypercube OK, that answers that question.
I guess I should make an appearance and partake of the free drinkies. Looks like I might be working from home on Thursday ;-D
23:48
@SimonRigharts Oh. Suddenly I am less keen.
@PaulWhite The presentations are still good, but I don't know how much you'd get out of them

« first day (782 days earlier)      last day (4383 days later) »