Well for the former its standard I/O -- if you wan...
# getting-started
r
Well for the former its standard I/O -- if you want to read the file in parallel you'll need to figure out stuff like record lengths and offsets, and then have your different threads reading the right bits. I wouldn't start with this though -- its unlikely to be your bottleneck. Reading a file synchronously is pretty fast if the file is not fragmented all over the disk.
1