why float no. become integer?
Asked by huda nawaf on 27 Apr 2012
Latest activity Commented on by Walter Roberson on 27 Apr 2012
hi,
i have two vectors
the first one is float as:
c=[2.5 4 6.5 4.5 3]; c1=[1 2 10 9];
when merge them as:
c2=[c c1]
why c within c2 become [2 4 6 4 3]?
thanks in advance
Products
No products are associated with this question.
3 Answers
Answer by Walter Roberson on 27 Apr 2012
Accepted answer
I bet if you check class(c1) you will find it is an integer data type.
2 Comments
Walter Roberson on 27 Apr 2012
Then the answer is "For reasons not explained, combining an integer data type in an array with floating point data is defined by MATLAB to return an integer data type."
If I were to speculate, then I would speculate that it is to allow people to write expressions such as
J = J + 1
when J starts as an integer data type here, it would be surprising for the user for it to suddenly become a floating point data type because the "1" happens to be floating point. It would be a nuisance for the programmer to have to continually write things like
J = J + ones(1, class(J));
Answer by Image Analyst on 27 Apr 2012
Because that's how MATLAB does it. If an integer is multiplied by a double, it gives a rounded integer result. Same thing if you combine them like you did. It does not promote the integer variable to the more general double like most languages, and like you would think - I don't know why, that's just the way it is. This was one of the most surprising things I learned when I was learning MATLAB. I can't find the explanation for this in the help - maybe someone else will say where it can be found.
1 Comment
Walter Roberson on 27 Apr 2012
When there is a mix of integer and floating point types, the result will be the left-most integer type evaluated in the expression:
>> ([int8([4 5 6]), uint16([1118 1119 11110])])
Warning: Concatenation with dominant (left-most) integer class may overflow other operands on conversion to return class.
ans =
4 5 6 127 127 127
0 Comments