Currently to read distributed seriesbinary data we use the spark sc.binaryRecords method. This depends on Hadoop, and may have unexpected ordering behavior across different systems. We should probably just write a method that maps the list of file names and reads the files directly, just as we read binary images.
Currently to read distributed
seriesbinarydata we use the sparksc.binaryRecordsmethod. This depends on Hadoop, and may have unexpected ordering behavior across different systems. We should probably just write a method that maps the list of file names and reads the files directly, just as we read binary images.cc @jwittenbach