Recently in learning to use a spark for the data mining community, the number of quantity is very big, in the call of RowMatrix computePrincipalComponents () method, the default always returns after calculation of Matrix, but my side operation data is too large, more than the biggest client receives memory, could you tell me what method can solve?

CodePudding user response:

In the form of block insert
Whether can consider

CodePudding user response:

And no,,, then find a temporary solution is to start the spark to augment the client's memory