Problems encountered during R's big data operation: there cannot be negative length Vectors

Solution:

You can only use trick to reduce the size of the generated intermediate matrix. For example, the problem I encountered is:

I want to run (w % * % H % * % T (h) (W is 5000000*10, and H is 10*100000)

The following error is returned when w % * % H is executed: A negative length vector is not allowed.

Google found that r documents and records related to Chinese characters were really poor. Finally, the English sentence "negative length vectors are not allowed" was used,

Only when we see the explanations from outsiders can we understand that, in short, the data size is too large, and the r system cannot generate results with the data size exceeding 2 ^ 31.

After changing to (w % * % (H % * % T (h), the problem is solved...