WPF Datagrid。(WPF Datagrid. Get the values of each cell of the selected row)

 除非您控制输入，否则最佳情况通常无用。 （即最好的情况通常是异常情况）。 除非它易于计算，否则不值得浪费你的时间。 
 
 平均情况：一般来说，这是你能想到的。 假设您使用大量输入，这通常是最有用的考虑因素。 
 
 最糟糕的情况：对于具有相同平均情况的两种算法，如果你处理任意输入（特别是如果它们是不可信的 - 即你接受来自网络上的人的输入），那么它是一个合适的打破平局。 在设计中也需要考虑一下 - 偶尔会出现这种情况。 一般来说，如果你有两个算法是O（n）平均情况，但一个是O（n lg n）最坏情况，一个是O（n ^ 2） - 它可能会影响你决定使用哪个。 或者它可能会影响您的算法设计。 
 
 示例：快速排序v。合并排序。 两者都是O（n lg n）Quicksort的最坏情况是O（n ^ 2），合并排序是（IIRC）仍然是O（n lg n） - 但是一般来说，如果数据适合内存，Quicksort往往会更快。 即使它有一个更昂贵的最坏情况，因为我们知道它有一个更昂贵的最坏情况，我们可以尝试缓解它（三个中位数而不是一个随机分区等）并利用它通常的事实比mergesort快。 

Best case is generally useless unless you control the inputs. (i.e. best case is usually an anomalous case). Unless it's easy to compute, it's not worth wasting your time.
 
Average case: it's what can you expect in general. Assuming you work with a large range of inputs, this is usually the most useful thing to consider. 
 
Worst case: a decent tie breaker for two algorithms with the same average case, if you deal with arbitrary inputs (especially if they're untrusted - i.e. you're accepting inputs from people on the web). Also something to consider in design in general - this will come up occasionally. In general, if you have two algorithms that are O(n) average case, but one is O(n lg n) worst case, and one is O(n^2) - it may influence your decision on which to go with. Or it may influence your algorithm design.
 
Example: quick sort v. merge sort. Both are O(n lg n) Quicksort's worst case is O(n^2), merge sort is (IIRC) still O(n lg n) - but in general, Quicksort tends to be faster if data fits in memory. Even though it has a costlier worst case, since we KNOW it has a costlier worst case, we can try to mitigate it (median-of-three instead of just a random partition, etc.) and take advantage of the fact that it usually is faster than mergesort.