Nearby Fulfillment Centers with Inventory

A delivery must be fulfilled near a destination center. The fulfillment network is an undirected graph whose edges are given by connections. Each row [u, v] connects centers u and v in both directions.
The array inventory represents a map from center ID to available quantity: each row is [centerId, quantity], and every center ID appears exactly once.
Return every center ID that satisfies all of these conditions:
the center is not destination;
its inventory quantity is greater than 0;
its shortest-path distance from destination is at most maxStep edges.
Return the qualifying IDs in ascending numerical order. A center in a disconnected component does not qualify.

Function
findFulfillmentCenters(connections: int[][], destination: int, maxStep: int, inventory: int[][]) → int[]

Examples
Example 1
connections = [[1,2],[1,3],[2,4],[3,4],[4,5]]
destination = 4
maxStep = 1
inventory = [[1,2],[2,0],[3,5],[4,3],[5,6]]
return = [3,5]
Centers 2, 3, and 5 are one edge from destination 4. Center 2 has no inventory, while centers 3 and 5 have positive inventory. The destination itself is excluded.
Example 2
connections = [[10,20],[20,30],[30,40]]
destination = 20
maxStep = 2
inventory = [[10,1],[20,9],[30,0],[40,4],[50,8]]
return = [10,40]
Center 10 is one edge away and center 40 is two edges away. Center 30 has zero inventory, and center 50 is disconnected.
Example 3
connections = [[1,2],[2,3]]
destination = 2
maxStep = 0
inventory = [[1,7],[2,5],[3,9]]
return = []
With maxStep = 0, only the destination is at an allowed distance, and the destination must be excluded.

Constraints
1 <= inventory.length <= 2 * 10^5.
Each inventory[i] is [centerId, quantity]; center IDs are distinct positive integers, and 0 <= quantity <= 10^9.
destination appears exactly once in inventory.
0 <= connections.length <= 2 * 10^5; each row contains two distinct center IDs that both appear in inventory.
0 <= maxStep <= 2 * 10^5.
The graph may be disconnected, and the answer is returned in ascending numerical order.

Build an adjacency list from connections, then run BFS from destination, tracking distance per node. Stop expanding once distance reaches maxStep. Collect any node that isn't the destination, has quantity above 0, and was reached within the limit. Sort the result ascending. The pitfall that sinks first attempts is the destination. With maxStep = 0 it's the only reachable node and it must be excluded, as Example 3 shows. Other traps: building the adjacency from inventory IDs that have no edges, so isolated centers like 50 never get visited, and using DFS, which gives wrong shortest distances. Use a hash map for inventory lookups since IDs aren't contiguous. Complexity is O(V + E + k log k) with up to 2 * 10^5 of each, so recursion depth is a risk too. Go iterative with a queue. If you freeze on the OA, StealthCoder can surface this BFS template while you verify the edge cases.

The honest play: practice the pattern, and have StealthCoder ready for the one you didn't see coming.