uniqx ver.1.0
date: 2005/02/03

uniq is buggy and too poor

usage:
	uniqx [-udcrf] [-k key,key,...] [file]

flags:
	-u 	# print only unique line
	-d	# print only duplicated line
	-c	# add count of duplicated line in printing
	-r	# print last line of duplicated line
	-f	# field wise comparison (default is column wise comparison)
	-k key,key,...

key:
	key specifies the fields to compare.
	key follows Python like syntax of range.
	n:m	# the meaning is n,n+1,...,m-1
	:m	# the meaning is 0:m
	n:	# the meaning is n to last
	:	# the meaning is 0 to last
	n	# the meaning is single field.
	n begins with 0

#
# Compatibility to uniq
#
I tried to support syntax of uniq, but it makes the code dirty.
I gave up.

the relation to uniq:
	uniq +n
is same as
	uniqx -k n:
and
	uniq -n
is same as
	uniqx -fk n:

#
# BUGS
#
Only tokenize is supported in getting fields.
Field separator option like
	-s separator
might be needed for special case.
However the meaning is unclear for me.
Therefore this option is left for future update.

Kenar (Kenji Arisaw)
E-mail: arisawa@aichi-u.ac.jp