I have an integration on the form $h(a)=\int \delta(a-g(x)) f(x) dx$, where f(x) and g(x) are numerical functions defined from the interpolation of experimental data (obtained from simulations).
I used the composite rule for $\delta(a-g(x)) = \sum_{i} \frac{\delta(x-x_i)}{|g'(x_i)|}$ to solve this integral.
Now, I have another set of data which results in $g'(x_i)=0$ at certain values of a, how can i go around it?
