Two dependent switch on float in a loop.

Question

Peiyuan ZHU 2017-5-13

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/340069-two-dependent-switch-on-float-in-a-loop

编辑： Jan 2017-5-13

for t=0.2:0.1:4
    fcn(t);
end
function fcn(t)
for x=1:3
    fcn1(x,t)
end
end
function fcn1(x,t)
switch t
    case 0.2
        y=3;
    case 0.3 % why is this skipped when t = 0.3?
        y=2;
    case 0.4
        y=1;
    otherwise
        error('t out of range');
end
switch y
    case 1
        z=1;
    case 2
        z=2;
    case 3
        z=3;
    otherwise
        error('y out of range');
end
z = z+x;
end

What's going on when t = 0.3? The switch case misjudges?

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Jan 2017-5-13

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/340069-two-dependent-switch-on-float-in-a-loop#answer_266892

编辑：Jan 2017-5-13

在 MATLAB Online 中打开

Wow, the most frequently asked question again.

Please read FAQ: why is 0.3-0.2-0.1 not equal 0 .

why is this skipped when t = 0.3

Because t is not 0.3 . You see that Matlab is convinced, that this case does not happen. Try it:

0.2 + 0.1 - 0.3
% >> 5.55111512312578e-17

Decimal numbers do not have an exact representation in the binary IEEE754 format, which is used for computations in Matlab and otehr programming languages. Therefore you get rounding effects due to the limited precision. See also:

1e17 + 1 - 1e17

When you want to perform numerical computations, it is very essential, that you understand what happens here and therefore this topic is handled in the first lesson and in the first chapter of books about numerics.