A natural number has pairwise distinct digits. Multiplying by gives a number whose digit sum is . How many such are there?
Turn the digit sum of into a count of borrows. Write and subtract in columns. In any subtraction, each borrow removes from one column and adds to the next, so it lowers the digit sum of the result by . Since and have the same digit sum,
where is the number of borrows in . (Checked on random values of : the identity holds every time.)
Read off how many digits must have. The condition becomes , i.e. exactly borrows. Borrows can only occur in the columns where actually has a digit, so where is the number of digits of . Hence
This is the step that makes the problem finite instead of open-ended.
Bound from above using the distinct-digit condition. Only ten digits exist, and they must all differ, so . Combined with the previous step, has either or digits — and a -digit with distinct digits must use every digit once, i.e. it is pandigital with .
Count the candidates in each case. For there are choices (choose and order nine of the ten digits, excluding a leading zero); for there are pandigital numbers. Roughly million candidates — small enough to test each one exactly.
Test them and collect the survivors. Exhaustive enumeration gives qualifying nine-digit number and qualifying ten-digit numbers:
The nine-digit one is , since with digit sum . The smallest ten-digit one is , since , again with digit sum .
Sanity-check the two extremes. For every one of the nine columns borrows, which is the maximum available and exactly the nine required. A tempting but wrong guess is , giving — its digit sum is right, but its digits are all equal, so it fails the distinctness condition. That near-miss is why the answer is not zero: the qualifying products look like , not like a string of nines.
Need to solve a different problem like this? Open the solver →