

Ten Centuries
Group: General Forum Members
Last Login: Monday, March 16, 2015 1:09 AM
Points: 1,263,
Visits: 1,082


tilew948340 (3/20/2011)
I am sorry, but I realy, realy don't understand why D is not good, even if I do your formula. Why is so that D would have a truncate answer and not C? I mean, both have a precision of 51 and only the scale is different (3 more digit more for C), which might explain something if you could explain my question here: If I declare D as precision of 23 instead of 25 and still having a scale of 10 DECLARE @value1D DECIMAL(23,10), @value2D DECIMAL(23,10) it is still a precision over 38, and the scale is still 20 as previous, but the answer is not truncate Why???
First of all a great big thank you to Duncan for this excellent QotD and the explanation.
Whether the decimal result is 'truncated' or not is a mere mathematical question: D would result in precision 51 and scale 20; in order to not truncate the integer part of the numeral, SQL Server does the following:  maximum precision = 38, desired precision is 51 ==> 51  38 = 13  since it doesn't truncate the integer part, the decimal portion (scale) is truncated: 20  13 = 7. Hence the result for option D is DECIMAL(38,7).
If you use a precsion of 23, the math is as follows:  Precision: 47  38 = 9  Scale: 20  9 = 11  Result: DECIMAL(38,11)
However, as Duncan stated in his explanation, scale will never be less than 6; so the 'minimum' result in regards to scale will always be DECIMAL (38,6).
Regards, Michael




SSCommitted
Group: General Forum Members
Last Login: Wednesday, May 27, 2015 3:44 AM
Points: 1,865,
Visits: 596


Good question, great answer
Thanks, Tom
Life: it twists and turns like a twisty turny thing




Ten Centuries
Group: General Forum Members
Last Login: Saturday, March 5, 2016 7:55 PM
Points: 1,354,
Visits: 1,300


Excellent question but this one got me. At first, I thought they would all return the same results and I was wrong so I definitely learned something new today.
Kudos to the author of this question.




Ten Centuries
Group: General Forum Members
Last Login: Thursday, December 31, 2015 5:26 AM
Points: 1,212,
Visits: 798


Great Question with even more great explanation. Good start for Monday




SSCertifiable
Group: General Forum Members
Last Login: Today @ 8:07 AM
Points: 7,713,
Visits: 11,071


Good question; superb explanation!
I guess that if the author had included "all of the above" as a fifth answer option, the rate of incorrect answers would have been a lot higher. I guess that is the answer most respondents will first have in mind.
Hugo Kornelis, SQL Server MVP Visit my SQL Server blog: http://sqlblog.com/blogs/hugo_kornelis




SSCrazy
Group: General Forum Members
Last Login: Saturday, September 17, 2016 8:01 AM
Points: 2,834,
Visits: 2,629


Excellent explanation, thanks.




Ten Centuries
Group: General Forum Members
Last Login: Saturday, March 5, 2016 7:55 PM
Points: 1,354,
Visits: 1,300


Hugo Kornelis (3/21/2011) Good question; superb explanation!
I guess that if the author had included "all of the above" as a fifth answer option, the rate of incorrect answers would have been a lot higher. I guess that is the answer most respondents will first have in mind. That would have been the option I selected. Or, if the question was one of those "Select All That Apply" and were check boxes instead of radio buttons I would have selected them all.




Say Hey Kid
Group: General Forum Members
Last Login: Wednesday, November 25, 2015 11:49 AM
Points: 671,
Visits: 489





Ten Centuries
Group: General Forum Members
Last Login: Friday, September 16, 2016 1:49 PM
Points: 1,265,
Visits: 13,613


fantastic question!!!
rfr.ferrari DBA  SQL Server 2008 MCITP  MCTS
remember is live or suffer twice! the period you fastest growing is the most difficult period of your life!




Mr or Mrs. 500
Group: General Forum Members
Last Login: Sunday, June 14, 2015 7:39 AM
Points: 559,
Visits: 1,248


Great question.
I must say, .NET Framework has one up on SQL here. Considering the following VB.NET code.
System.Console.WriteLine(1234567890.123456789D * 0.1D * 0.1D) returns 12345678.90123456789 as expected which is simple.
DECLARE @value1 DECIMAL(20,10), @value2 DECIMAL(2,1) , @value3 DECIMAL(2,1) Will return 12345678.901234567890 as expected
But if you make a mistake and just use the same data type for all... DECIMAL(20,10) then it breaks. 12345678.901235
I guess we now know why we don't have a Product aggregate function in SQL.



