How to find common strings among two very large files?

Question

I have two very large files (and neither of them would fit in memory). Each file has one string (which doesn't have spaces in it and is either 99/100/101 characters long) on each line.

Update: The strings are not in any sorted order.
Update2: I am working with Java on Windows.

Now I want to figure out the best way to find out all the strings that occur in both the files.

I have been thinking about using external merge sort to sort both the files and then do comparison but I am not sure if that would be the best way to do it. Since the strings are mostly around the same length, I was always wondering if computing some kind of a hash for each string would be a good idea, since that should make comparisons between strings easier, but then that would mean I have to store the hashes computed for the strings I have encountered from the files so far so that they can be used later when comparing them with other strings. I am not able to pin down on what exactly would be the best way. I am looking for your suggestions.

When you suggest a solution, also please state if the solution would work if there were more than 2 files and strings which occur in all of them had to be figured out.

Answer 1

I would sort each file, then use a Balanced Line algorithm, reading one line at a time from one file or the other.

Answer 2

You haven't said what platform you're working on, so I assume you're working on Windows, but in the unlikely event that you're on a Unix platform, standard tools will do it for you.

sort file1 | uniq > output
sort file2 | uniq >> output
sort file3 | uniq >> output
...
sort output | uniq -d

Answer 3

I'd do it as follows (for any number of files):

• Sort just 1 file (#1).
• Walk through each line of the next file (#2) and do a binary search on the #1 file (based on the number of lines).
• If you find the string; write it on another temp file (#temp1).
• After you finished with #2, sort #temp1 go to #3 and do the same search but this time on #temp1, not #1, which should take much less than the first one as this only has repeated lines.
• Repeat this process with new temporary files, deleting previous #temp files. Each iteration should take less and less, as the number of repeated lines diminishes.