Another fallacy of equivocation.
The set I am talking about is the set of digits that is drawn from for this infinite sequence. The set itself is not infinite. Set A has 10 members {0,9}. Set B has 9 members {0,2-9}.
You are arguing based on a different understanding of what the sets are. Even still, your argument does not hold up.
In your case, let set X be the set of all infinitely long number randomly generated. Pick any number from this set, and remove all of the ones. The resulting number exists in a set that is not a subset of set X. Why? Because set X is the set of all numbers that are both infinitely long and random and it is impossible for a number without 1s to exist in set X. Both conditions are important.
If the number is truly randomly generated, each digit (0-9) have an equal probability of appearing. For a finite length of digits (even if said length is large), it is possible (though improbable) that the number wouldn't contain a 1. If we were arguing for finite length numbers, your argument would be fine. However, for infinite length, it is impossible because as length approaches infinity, the probability of a digit not appearing approaches 0. At infinity, it would be 0. This is where your argument breaks down. The number with all 1s removed is not a part of set X so your argument above as a counter example is invalid. This number would be a part of set Y where Y is the set of all infinitely long numbers randomly generated from set B.
These sets are not equivalent, and it is important to keep this in mind when constructing a proof.