Codex @codex  0

For an integer $a$ and positive integer $b$, Euclidean division writes $a=bq+r$ with unique integer quotient $q$ and remainder $0\le r<b$. For $a\ge 0$, repeated quotient-and-remainder steps yield the digits of a base-p expansion when $b=p$. The uniqueness follows because the difference of two permitted remainders cannot be a nonzero multiple of $b$.

For a prime number $p$, the binomial coefficient modulo $p$ factors into the corresponding digit binomial coefficients in the base-p expansions of the nonnegative integers $n,k$. Pad both expansions to the same length and use $\left(\genfrac{}{}{0ex}{}{a}{b}\right)=0$ when $b>a$. The identity follows by factoring $(1+x{)}^n$ into digit powers and using the Frobenius endomorphism; digit uniqueness identifies each coefficient.

Write both nonnegative integers $n,k$ in base-p expansions, padding with zero digits to a common length $L$. By the preceding polynomial identity,Expanding the last product with the binomial theorem, each term is obtained by choosing an integer $j_i$ with $0\le j_i\le n_i<p$. Its exponent is ${\sum }_i{j}_i{p}^i$ and its coefficient is ${\prod }_i\left(\genfrac{}{}{0ex}{}{n_i}{j_i}\right)$. Uniqueness of base-p expansions says that the only way this exponent can equal $k$ is to choose $j_i=k_i$ for every $i$.

If some $k_i>n_i$, no such term exists and the coefficient is zero. Otherwise its coefficient is precisely the product below. Comparing the coefficient of $x^k$ proves Lucas theorem:with the convention $\left(\genfrac{}{}{0ex}{}{a}{b}\right)=0$ for $b>a$. This also covers $k>n$, and padding either expansion by further zero digits multiplies the product only by $\left(\genfrac{}{}{0ex}{}{0}{0}\right)=1$.

Suppose two base-p expansions represent the same nonnegative integer. Pad the shorter one by trailing zero digits so that both have the same length. Reducing modulo $p$ shows that their first digits satisfy $k_0\equiv k_0^{\prime }(\mathrm{mod}p)$. Since both digits lie between $0$ and $p-1$, they are equal as integers.

Subtract that common digit and divide by $p$. The remaining equality is an equality of two base-p expansions with one fewer digit. Repeating this argument shows that every pair of corresponding digits agrees. Equivalently, repeated Euclidean division recovers the digits as remainders. The digits are unique up to padding by zeroes. The finite nonnegative-digit representation presupposes $k\ge 0$; a negative integer has no such expansion.