Hi Yanbo, I'm using java language and the environment is spark 1.4.1. Can u tell me how to do it more detail , the follows is my code, how can i save the cov to hdfs file ? " RowMatrix mat = new RowMatrix(rows.rdd()); Matrix cov = mat.computeCovariance(); " The cov is localMatrix, if the Matrix is much more than 5G , is there any better way to do it. THX. ------------------ ???????? ------------------ ??????: "Yanbo Liang";<yblia...@gmail.com>; ????????: 2016??1??25??(??????) ????1:31 ??????: "zhangjp"<592426...@qq.com>; ????: "user"<user@spark.apache.org>; ????: Re: how to save Matrix type result to hdfs file using java
Matrix can be save as column of type MatrixUDT.