Round Prices to Match Target

Given a float array prices and an integer target, round every price to either its floor or its ceiling so that the rounded values sum to target while minimizing the total absolute rounding error.
For example, prices = [1.2, 4.3, 5.8, 6.4] and target = 18 returns [1, 4, 6, 7].
Practice rule
If equal fractional parts compete for the last round-up position, prefer the smaller original index.
If no floor-or-ceiling assignment can sum to target, return an empty array.

Function
roundPricesToMatchTarget(prices: float[], target: int) → int[]

Examples
Example 1
prices = [1.2, 4.3, 5.8, 6.4]
target = 18
return = [1, 4, 6, 7]
:)

Floor everything first and add up the floors. The gap between target and that sum is how many prices must round up. If the gap is negative or bigger than the count of prices with a nonzero fractional part, no assignment works, so return an empty array. Otherwise, sort indices by fractional part descending, breaking ties by smaller original index, and round up the top k. That minimizes error because rounding up a price costs (1 - frac) instead of frac, so the biggest fractions are the cheapest to bump. The pitfall is whole-number prices. Their floor and ceiling are equal, so they can't absorb a round-up and shouldn't count toward the gap. Float precision bites too, so compute fractions carefully. Sorting makes it O(n log n). During the live OA, StealthCoder is the hedge if the tie-break or the infeasible case slips your mind.

Drill it cold or hedge it with StealthCoder. Either way, don't walk into the OA hoping you remember the trick.