问题
I am extracting cells out of a spreadsheet using the Interopt.Excel API. When I call:
object[,] rangeValues = (object[,])range.get_Value(XlRangeValueDataType.xlRangeValueDefault);
And set a breakpoint and inspect rangeValues, I see elements starting at [1,1] "foo", [1,2] "bar", etc.
However, if I do string[,] test = new string[2, 2] { { "one", "two" }, { "three", "four" } };
The elements start at [0,0]. How does the Excel API construct a multidimensional array w/ empty elements? I tried adding null but you still have an [0,0] entry. Their object doesn't show that.
回答1:
The CLR supports non-zero based arrays. They are confusing and to be avoided. I believe, the COM interop marshaller can create them.
http://msdn.microsoft.com/en-us/magazine/cc301755.aspx
Common Language Specification (CLS) compliance requires that all arrays be zero-based. This allows a method written in C# to create an array and pass the array's reference to code written in another language such as Visual Basic®. In addition, since zero-based arrays are by far the most common, Microsoft has spent a lot of time optimizing their performance. However, the CLR does support non-zero-based arrays but they are discouraged. For those of you who do not care about performance and cross-language portability, I will demonstrate how to create and use non-zero-based arrays later in this section.
This is the first time I have seen one in real-world code.
Apparently, multi-dimensional arrays have the same type no matter whether they are zero-based or not. Single-dimensional ones have a different type. Makes sense because if they had the same type the JIT would always have to produce slow code for the general case.
来源：https://stackoverflow.com/questions/23893343/non-zero-based-multidimensional-arrays