How should I practice Optimal Transfer?

Start by reading the full statement, solve the visible examples, then compare the pattern against related Snowflake online assessment problems in FastPrep.

Is Optimal Transfer associated with Snowflake?

FastPrep tags this as a Snowflake online assessment coding practice problem when the problem catalog has that company and stage context.

Optimal Transfer

Snowflake● MediumINTERNOA

Learn

Problem statement

There are n servers indexed from 0 to n - 1. Their capacities are given in a strictly increasing integer array capacity.

The distance between servers x and y is |capacity[x] - capacity[y]|. Each server has a unique closest other server by this distance.

From a server x, you may make either type of move:

Move directly to any other server y for a cost of |capacity[x] - capacity[y]|.
Move to the closest server of x for a cost of 1.

The fixed-cost move is directed: if y is closest to x, x is not necessarily closest to y. You may use any number of moves, and a path costs the sum of its moves.

For every query i, find the minimum cost to travel from server fromServer[i] to server toServer[i]. Queries are independent and do not change any server. Return the costs in query order. When the starting and destination indices are equal, the empty path costs 0.

Function

minimumTransferCost(capacity: int[], fromServer: int[], toServer: int[]) → int[]

Examples

Example 1

capacity = [2,7,10]fromServer = [0,1,2]toServer = [2,2,1]return = [2,1,1]

The closest-server moves are 0 → 1, 1 → 2, and 2 → 1, each costing 1. The first query uses 0 → 1 → 2 for a total of 2; the other two each use one closest-server move.

Example 2

capacity = [1,4,9,15]fromServer = [0,3,1]toServer = [3,0,3]return = [12,3,11]

The closest-server moves are 0 → 1, 1 → 0, 2 → 1, and 3 → 2. Traveling right from 0 to 3 costs 1 + 5 + 6 = 12. Traveling left from 3 to 0 costs 1 + 1 + 1 = 3. The last query costs 5 + 6 = 11.

Example 3

capacity = [5,6,20]fromServer = [2,0,1]toServer = [0,1,1]return = [2,1,0]

The first query follows 2 → 1 → 0 for a cost of 2. Moving from 0 to 1 costs 1. The last query stays at server 1, so it costs 0.

Constraints

For this exercise, assume 2 <= capacity.length <= 100000 and 1 <= capacity[i] <= 10^9.
capacity is strictly increasing, and every server has exactly one closest other server. In particular, an interior server never has equal left and right capacity gaps.
For this exercise, assume 0 <= fromServer.length == toServer.length <= 100000.
Each query index lies in [0, capacity.length - 1].
All minimum costs fit in a signed 32-bit integer under these bounds.