two way merge sort for 2 file c++ - c++

Question

I was asked to apply Two way merge sort on two files(files of records) ,
the algorithm explains the steps as follows :
Sort phase
1)The records on the file to be sorted are divided into several groups.
Each group is called a run, and each run fits into main memory.
2)An internal sort is applied to each run, 3)and the resulting sorted runs are distributed to two external files.
Merge Sort: 1) One run from each of the external files created in the sort phase merge into a larger runs of sorted records.
2)The result is stored in a third file.
3)The data is distributed back into the first two files, and merging continues until all records are in one large run.
I was able to apply Sort Phase only , so the current files is :
(supposed run contains 3 keys only )
file 1:
50 95 110 | 40 120 153 | 22 80 140
file 2:
10 36 100 | 60 70 130
here's the steps of merge phase
so if i will solve it theoretically will perform the following :
Merge Phase:
step1 :
file 3 :
10 36 50 95 100 110 | 40 60 70 120 130 153 | 22 80 140
file 1:
10 36 50 95 100 100 | 22 80 140
file 2 :
40 60 70 120 130 153
step 2 :
file 3 :
10 36 40 50 60 70 95 100 110 120 130 135 | 22 80 140
file 1 :
10 36 40 50 60 70 95 100 110 120 130 135
file 2:
22 80 140
step 3 :
file 3 :
10 22 36 40 50 60 70 80 95 100 110 120 130 135 140
one run stop sort complete
Now i need to apply merge phase so each key from each file compared to each other and output the smaller to file 3 , and in step 2 redistribute file 3 into two file then merge and sort until have one sort run
How i can apply such algorithm in c++ , i'm little bit confused about how can i determine the size of each run in every step.

Answer 1

As commented by Amdt Jonasson, the program needs to keep track of the run sizes and the end of data for each file. In your example it appears the initial run size is a fixed run size of 3 elements. A merge of two runs of size 3 will result in a single run of size 6 as shown in your steps. In this case, only a single instance of run size and the end of data in each file needs to be tracked.
If the sort is to be a stable sort (the original order preserved on equal keys), and the run size is variable, then an array of run counts for each file will be needed, or some way to denote the end of runs in the file, such as a text file, with a special character sequence used as a end of run indicator.
If the sort is not required to be stable, then an out of order sequence (smaller key value after a larger key value) can be used to indicate the end of a run. The risk here is that two or more runs will appear to be a single run if the runs happen to be in order, which will lose stability and unbalance the run count on the files.
This is a two way merge sort using 3 files. If you use a 4th file, then each merge of runs can alternate between 2 output files, eliminating the need split up the runs after each merge pass.
An alternative for doing a 2 way merge sort with 3 files is a polyphase merge sort, but it's complicated, probably beyond what would be expected for a class assignment, and more of a "legacy" algorithm used back in the days of tape based sorts.