why uniq don't give non-duplicated results

Question

find ./2012 -type f | cut -d '/' -f 5 | uniq

The usual filenames look like

./2012/NY/F/Zoe
./2012/NJ/M/Zoe

I suppose the command above should give non-duplicated result of file names like Zoe only for once, but it turns out not so.

Why? and how should I write to get the desired result?

Answer 1

uniq only detects duplicates if they're in consecutive lines. The usual idiom is to sort | uniq to ensure that any duplicates will appear together.

Answer 2

uniq requires the duplicates to be adjacent, which means you need to sort the input, which means you might as well use sort -u;

find 2012 -type f | cut -d/ -f5 | sort -u