Notice that we can find a good N such that after knowing d(Nx) then x can uniquely be determined. For this reason, we shall provide two different approaches;
1st approach. Let p1<p2<⋯<pt be all the primes less than 100! such that p1R>100!. Now, for each i, choose large enough primes qij and 1≤j≤R and put N=p1α1⋯ptαt such that ai≡−j(modqij), j=1,…,R. Now, if d(Nx) is divisible by qij it follows that vpi(x)=j and this would help us to uniquely determine x.
2nd approach. Let us denote by p1,…,pt the prime divisors dividing x. Then, x=∏i=1tpiαi, αi≥0. Then if N=∏i=1tpiβi it follows that d(Nx)=∏i=1t(1+αi+βi). Letting βi=x2i−1. Considering P(x)=∏i=1t(1+αi+x2i) then, degP(x)=2t−1 and the coefficient of x2t−2j−1 is equal to 1+αj. Take n>∏i=1t(1+αi) then P(n) will be a number in base n that all of its digits can uniquely be determined by αi. So, it remains to choose, n>∏i=1t(1+ai) where ai=maxi∈Aαi and βi=n2i−1.
Now, putting T=100!N and then d(φ(Tx))=d(φ(100!))Nx. Plugging m=pS−1N where t=q∏k=1100!d(kN) for some sufficiently large primes p and q. Then
φ(d(my))=(q−1)d(Ny)φ(k=1∏100!d(kN))
It follows that if we consider φ(d(my))+d(φ(Tx))(modq−1) we can determine d(φ(Tx)) and then determine x.