I was in the need of a unique word which, given a sequence, builds a
new sequence of the same kind with unique elements of the input
sequence while keeping the same order (only subsequent duplicate
elements must be ommited). I ended up with:
: unique ( seq -- newseq )
[
dup length
Hi Sam,
Take a look at the 'prune' word in the hashtables vocabulary, it does
exactly what you want but it runs in O(n) time because it uses a
hashtable to track which elements have already been added.
Slava
On 6-Oct-07, at 1:48 PM, Samuel Tardieu wrote:
I was in the need of a unique word