Break the problem into three parts:
1. Select i digits from nums1, k-i digits from nums2.: Try all valid splits.
2. Get max number of length m from one array.: Use monotonic stack (similar to Remove K Digits).
3. Merge two arrays into max number.: At each step, pick the lexicographically larger suffix. Combine: try all splits, merge each pair, keep the maximum result. The lexicographically larger suffix comparison in merging means: if nums1[i:] > nums2[j:] as arrays, pick from nums1. This tie-breaking rule ensures maximum result when both arrays have the same leading digit.