Hi, how can I use BigInt/Long as type of size in Vector.sparse?

[吴承霖]

Hi all,


I have to deal with a lot of data, and I use spark for months.

Now I try to use Vectors.sparse to generate a large vector of features, but the 
feature size may exceed 4 billion, above max of int, so I want to use BigInt or 
Long type to deal with it.

But I read code and document that Vectors.sparse support int as size type only, 
why limit it to int? And how can I bypass this limit?

Thanks,
Alexander