idx=find_idx(xi, xgrid)
This function not only bin similarly to HISTC, but also returns the fractional position of data points within the binning interval.
It is equivalent to
interp1(xgrid,(1:length(xgrid)), xi)
but with the speed improvement up to 5 times.
Algorithm: dichotomy, complexity of m.log(n), where m isnumber of data points (xi) and n is number of bins (xgrid).
Few obvious examples of applications:
 binning step for more sophisticated interpolation schemes such as multidimensional spline or linear tensorial interpolations.
 Generate discrete random sequences with given probability distribution.
