How to read 64 bytes each binaries into RDD in pyspark and do the filter?

Question

I want to read a binary file into RDD. Reading 64 bytes each time(but the number of bytes at the end of the file might not be full 64-byte). How can I do it? And how to filter