There are two ordered list, the list a and the list b, such as:
a=>[1,14,20,36]
b=>[2,7,11,25,32,39]
the result that I want to get is:
result=>[(2,1),(7,null),(11,14),(25,null),(32,null),(39,36)]
Put it into word: find the closest number(which delta between counterpart is less than 5) in list a for each item in list b.
First of all, I decide to divide the problem to:
1,find closest number for each item in a, since a is less elements than b.
2,compose a result into a list. if no proper value be found, put it to null.
Because a&b are ordered list, so binary search is fast way to find out corresponding closest value. The complexity of binary search for a ordered list is O(log(N)).Each of item in list a have O(N), so the total complexity is O(N*log(N))。
How to use the code:
static void Main(string[] args) { int[] arrTime = { 1, 3, 5, 7, 8, 9, 10, 12, 15, 18, 25 }; int[] arrAlign = { 3, 6, 11, 15, 22 }; FindClosestNumber finder = new FindClosestNumber(3); var lst = finder.FindClosestAlignment(arrTime, arrAlign); foreach (var item in lst) { Console.WriteLine("{0}\t\t{1}", item.Time, item.Alignment); } Console.ReadKey(); }
There is the code:
public class Pair { public Pair(int? time, int? alignment) { Time = time; Alignment = alignment; } public int? Time { set; get; } public int? Alignment { set; get; } } public class FindClosestNumber { public int Threshold { get; set; } public FindClosestNumber() { this.Threshold = 50; } public FindClosestNumber(int threshold) { this.Threshold = threshold; } public List<Pair> FindClosestAlignment(int[] arrTime, int[] arrAlignment) { List<Pair> list = FindClosestTime(arrTime, arrAlignment); List<Pair> reslut = new List<Pair>(); for (int i = 0; i < arrTime.Length; i++) { Pair pair = new Pair(arrTime[i], null); foreach (var p in list) { if (pair.Time == p.Time) pair.Alignment = p.Alignment; } reslut.Add(pair); } return reslut; } private List<Pair> FindClosestTime(int[] arrTime, int[] arrAlignment) { List<Pair> list = new List<Pair>(); for (int i = 0; i < arrAlignment.Length; i++) { int el = arrAlignment[i]; int index = FindClosestIndexInOrderList(el, arrTime, this.Threshold); if (index == -1) { list.Add(new Pair(null, el)); } else { list.Add(new Pair(arrTime[index], el)); } } return list; } private int FindClosestIndexInOrderList(int val, int[] arr, int threshold) { int index = Array.BinarySearch(arr, val); if (index >= 0) {//found return index; } else { //bitwise complement for index int compl = (~index); if (compl == arr.Length) {//val is biggest value in the arr if (Math.Abs(arr[arr.Length - 1] - val) > threshold) return -1; else return (arr.Length - 1); } else if (compl == 0) {//val is smallest value in the arr if (Math.Abs(arr[0] - val) > threshold) return -1; else return 0; } else { int preIndex = compl - 1; int deltaToPrevious = Math.Abs(arr[preIndex] - val); int deltaToCurrent = Math.Abs(arr[compl] - val); if (deltaToPrevious > threshold) return -1; if (deltaToCurrent > threshold) return -1; return deltaToPrevious <= deltaToCurrent ? preIndex : compl; } } } }