October 17, 2009 at 5:25 pm
Comments posted to this topic are about the item A Wish list for the Microsoft Flying Squad
October 17, 2009 at 10:06 pm
Hi Tony,
Funny you mentioned extended properties π I'm working on a little set of apps right now that will greatly simplify working with them. Of course the Intellisense won't be modified to incorporate them, but adding, deleting, updating and retrieving them on objects will be greatly simplified. To me it only makes sense to let the server manage your metadata, but as you mentioned it can be painful to use the SPs.
October 18, 2009 at 12:11 am
Another useful feature would be change in versioning so it accepts the same plugins that Visual Studio uses. For now versioning doesn't work for most people around.
October 18, 2009 at 3:51 am
I realize that this is not a major issue, but I always find it irritating that the system stored procedures haven't been updated with the new data types. ---for years and years and years.....
Try this...
[font="Courier New"]DECLARE @Output VARCHAR(8000)
EXECUTE xp_sprintf @output out,
'Go hang yourself, you old M.D,!
You shall not sneer at me.
Pick up your hat and stethoscope,
Go wash your mouth with laundry soap;
I contemplate a joy exquisite
In not paying you for your visit.
I did not call you to be told
My malady is a common cold.
By pounding brow and swollen lip;
By fever''s hot and scaly grip;
By those two red redundant eyes
That weep like woeful %s skies;
By racking snuffle, snort, and sniff;
By handkerchief after handkerchief;
This cold you wave away as naught
Is the damnedest cold man ever caught!', 'April'
SELECT LEN(@Output)
--253
[/font]
Yup, the documentation proudly says that it takes a string. ...Er... only 256 characters. Look here, this isn't even lined up for deprecation so how long would it take to fix to make it use MAX??
Best wishes,
Phil Factor
October 18, 2009 at 6:36 am
While table partitioning has been available since 2005, it's almost easier to continue to manually manage it with regular tables, since tables have to be swapped in and out, and partition functions changed with a lot of verbose code - i.e. equivalent to the same code you would change when you added partitioning tables into a view and changed the trigger which makes the view updatable.
October 18, 2009 at 6:37 am
If they were to fix extended properties, they should fix it so that they are not lost when altering a multi-step table value function whenever I remove schemabinding and re-add schemabinding.
October 18, 2009 at 10:06 am
Good editorial. I'm not sure what I'd want fixed more, though I think partitioning for sure could use an easy way to manage it (and then generate the scripts).
I might ask for more monitors for service broker and mirroring as well.
October 18, 2009 at 12:20 pm
In most software development projects there is a bug I call the sycophant who tells the boss he or she can do the job and most bosses usually find out after the person with the implementation detail have left without leaving the implementation detail. I can tell you the people who created those nice features have moved on to other companies Microsoft is left without implementation detail, Microsoft is also in a better position than other companies because their employees must leave blogs as is. If the employee did not blog about it the feature is usually lost forever even with a blog I think Microsoft sometimes hire 5 to 10 new people who still cannot complete the nice feature. I know one such work in .NET I help their customers only as needed because the person who created it left with the implementation detail. I was surprised to find the personβs blog number one when I searched Google for it using the feature as key word because I did not remember his name , who knows if he now works for Google.
:Whistling:
π
Kind regards,
Gift Peddie
October 18, 2009 at 6:38 pm
Gift Peddie's good post brings up a real sore spot with me... why does everyone think that the Agile Manifesto says "[font="Arial Black"]No[/font] Documentation" whatsoever? Why does everyone think that if you document code (especially embedded documentation) that you're a sissy of some sort? Why does everyone think that having a simple implementation plan in black and white is somehow against the ability to move quickly?
IT managers of the world... wake up. Insist on proper embedded documentation so it never gets separated from the code and then give people time to do it right. Then insist on code reviews and make them part of everyone's job using pay reviews as enforcement. If you lose a couple of programmers because of it, good... they needed to be replaced anyway.
And remember... "If you want it real bad, you'll probably get it that way." π
--Jeff Moden
Change is inevitable... Change for the better is not.
October 18, 2009 at 7:10 pm
IT managers of the world... wake up. Insist on proper embedded documentation so it never gets separated from the code and then give people time to do it right. Then insist on code reviews and make them part of everyone's job using pay reviews as enforcement. If you lose a couple of programmers because of it, good... they needed to be replaced anyway.
And remember... "If you want it real bad, you'll probably get it that way."
This is more about intellectual property transfer than documentation because I can give you ten pages of documentation without the three lines needed to implement what I am working on. It takes very little to keep a developer happy but companies sometimes have paid team members doing nothing, writing useless compiled code, working on manager's pet project reinventing the wheel. I find most large projects are useless without a few people in the team and if these people are marginalized so they move on the project is dead.
Kind regards,
Gift Peddie
October 18, 2009 at 8:37 pm
Agreed and that's part of my point. The managers need to stop trivializing and take control/ownership of their projects. If work is lost just because an employee (at any level) moves on to another opportunity, then the company isn't doing things right.
--Jeff Moden
Change is inevitable... Change for the better is not.
October 19, 2009 at 6:48 am
PIVOT is definitely one unfinished or not properly finished feature. You need to hard code the values for the pivoted column. Otherwise you still to use dynamic SQL. E.g.:
USE AdventureWorks;
GO
SELECT VendorID, [164] AS Emp1, [198] AS Emp2, [223] AS Emp3, [231] AS Emp4, [233] AS Emp5
FROM
(SELECT PurchaseOrderID, EmployeeID, VendorID
FROM Purchasing.PurchaseOrderHeader) p
PIVOT
(
COUNT (PurchaseOrderID)
FOR EmployeeID IN
( [164], [198], [223], [231], [233] )
) AS pvt
ORDER BY VendorID
It does not support a subquery in the IN clause.
October 19, 2009 at 8:25 am
Agree with Peter on Pivot: totally useless function as implemented, would be useful if it didn't require hard-coding.
Agree with Gift and Jeff on documentation and "passing the baton". Any project that's in a position where losing a person, either due to employment termination or accident/illness, needs a better manager.
- Gus "GSquared", RSVP, OODA, MAP, NMVP, FAQ, SAT, SQL, DNA, RNA, UOI, IOU, AM, PM, AD, BC, BCE, USA, UN, CF, ROFL, LOL, ETC
Property of The Thread
"Nobody knows the age of the human race, but everyone agrees it's old enough to know better." - Anon
October 19, 2009 at 8:55 am
Agree with Gift and Jeff on documentation and "passing the baton". Any project that's in a position where losing a person, either due to employment termination or accident/illness, needs a better manager.
People moving on is not related to illness or termination more like let your sycophants do the work for you and this affects for than 40 percent of software development projects. These are not isolated manager should be replaced issues marginalized people move on at least one or more everyday in the US that is the reason 60 percent software projects fail.
Microsoft is in a better position to mitigate these issues more than other companies because other people can use what is more like the public documentation of the feature in most cases code included.
Kind regards,
Gift Peddie
October 19, 2009 at 9:22 am
GSquared (10/19/2009)
Agree with Peter on Pivot: totally useless function as implemented, would be useful if it didn't require hard-coding.
If it can produce variable number of columns depending on data, its output schema is unknown until run time - so it may as well build it with dynamic SQL anyway. I agree there needs to be an easier way to build them, but any alternative syntax will still have this fundamental problem.
Viewing 15 posts - 1 through 15 (of 17 total)
You must be logged in to reply to this topic. Login to reply