best performance ?

[Edward Chen]

Assume I have a txt file containing ten millions phone number, record unsorted and duplicate, Now I want

1. list top 20 duplicate phone numbers
2. sorted it
3. list duplicate frequency, like one phone number has 200 duplicate.

Which way has best performance? Database is not in the option list.

[William Brogden]

1. devise a way to turn the text of a phone number into a Java primitive, probably a long,
2. scan list adding the derived longs to a long[] array
3. sort the array

the remainder should be obvious.

Bill

[steve souza]

If it is in a file already you may not need to use java. You could also consider using unix utilities to sort and check for dupes.