I'm trying to prove the following result: If $f : [0,1] \to \mathbb{R}$ is integrable and the integral of $f$ is $0$ on every subinterval $I$ of $[0,1]$, then the integral of $\int_0^1|f|$ is $0$.
I have no idea how to proceed. Can you please suggest a method of attack or provide a proof?
Thank you.
In the following reasoning I have assumed that $f$ is continuous, the general case is put in the last paragraph, but it uses some Lebesgue integral theory.
Assume that $\displaystyle\int_{0}^{1}|f|>\epsilon_{0}>0$, then bisect $[0,1]$ and pick an interval $[a_{1},b_{1}]$ such that $\displaystyle\int_{a_{1}}^{b_{1}}|f|>\epsilon_{0}$, and bisect again to $[a_{1},b_{1}]$ to pick an interval such that...
Consider $\displaystyle\bigcap_{n}[a_{n},b_{n}]$, pick a $c$ of the intersection. If $f(c)<0$, choose a neighbourhood of $c$ such that $f$ has the negative sign around that neighbourhood, then by picking an interval $[a_{N},b_{N}]$ which is contained in the neighbourhood, you are now ready to get a contradiction.
For general case, here's a guess. We know that $f$ is continuous a.e. For such a continuous point $\alpha$, if $f(\alpha)>0$, then $f$ is positive around a neighbourhood of $a$, then the integral of $f$ around that neighbourhood must strictly greater than zero, but the assumption says that the integral must be zero, so $f=0$ for all continuous points. Now $\displaystyle\int|f|=0$ in the Lebesgue sense, but $|f|$ is Riemann integrable, then the Lebesgue integral and Riemann integral coincide.