file_sorting/TODO.md

• file operations

  • write to file (at index)
  • read file

• a tape type - representing a file?

• a record type - x1,x2,x3,x4,x5

  • comparing records

• buffer quick sort

• configuration

  • cmd args - man getopts

• file input

• keyboard input

• stage 1

  • reading chunks
  • sorting
  • writing runs

• stage 2

  • implement a min heap with following operations

    • pop
    • push

  • step 4

    • split the buffer into chunks of size n, and (b-1)n
    • create a min heap of size (b-1)n
    • populate the heap with (b-1) runs of n elements (each from a different file)
    • find a way to mark if a last element from a certain file has been popped
    • when a merged buffer is full, write it to file
    • pop off the top element until there are no more elements from a certain array
    • load another run from that file, and rehapify
    • repeat untill all files empty
    • the resulting file should get the new smallest filename in tmp dir

  • multiway merge

  • writing larger runs

  • repeat untill done

• counting read, and write operations

• testing

• report