Modulo with expression not the same as modulo with value

Why does mod(1, Y-X) not approximately equal 0?

mod(1, 0.1)

ans = 0

X = 2;
Y = 2.1;
Y - X

ans = 0.1000

mod(1, Y - X)

ans = 0.1000

I realize that Y - X ~= 0.1 but

isapprox(Y - X, 0.1, "loose")

ans = logical
   1

therefore, why not

isapprox(mod(1, Y - X), 0, "loose")

ans = logical
   0

My guess is that it is either something to do with floating point precision or the fact that it is an expression vs. a value?

Answer 1

[Matt J]

Because mod(1,z) is a discontinuous function of z at 0.1,

fplot(@(z) mod(1,z),[0,0.2])

●

[James]

Ah right, thought I was missing something obvious, thanks.