Get unique string occurrence and display it - awk

Question

I'm not really good with awk, so here what I just did to count the number occurence in one row:
the input.txt has this:
18 18 21 21 21 21 18 21
I just want to display the unique number that occur above. So, here is my code:
input="input.txt"
output=$(fmt -1 "$input" | sort | uniq | awk '{printf $1","}')
echo "$output"
The output:
18,21,
I got the result correctly but that comma , at the end, how do I remove that comma? Also, is there a simpler or a clean method without using fmt ?
The expected output:
18,21
Edit to remove comma, I use this:
sed 's/,$//'
and it's working, but is there a simpler way to do this without using fmt ?

Answer 1

Could you please try following.
awk '
BEGIN{ OFS="," }
{
for(i=1;i<=NF;i++){
if(!arr[$i]++){
val=(val?val OFS:"")$i
}
}
print val
val=""
}' Input_file
Explanation: Adding detailed explanation for above.
awk ' ##Starting awk program from here.
BEGIN{ OFS="," } ##Setting output field separator as comma here.
{
for(i=1;i<=NF;i++){ ##Traversing through all fields of currnet line here.
if(!arr[$i]++){ ##Checking condition if arr is NOT having current field present in it
val=(val?val OFS:"")$i ##Creating val and keep adding values to it, to print at last all values.
}
}
print val ##printing val here.
val="" ##Nullify val here.
}' Input_file ##mentioning Input_file name here.

Answer 2

Here is an alternative way in gnu awk:
awk -v RS='[[:blank:]]+' '!seen[$1]++{s=s (s!=""?",":"") $1} END{print s}' file.txt
18,21

Answer 3

Newer versions of perl has uniq in standard library. Otherwise, you'll have to manually write the logic (How do I print unique elements in Perl array?) or use https://metacpan.org/pod/List::MoreUtils
perl -MList::Util=uniq -lane 'print join ",", uniq #F'
perl -lane 'print join ",", grep { !$seen{$_}++ } #F'
With ruby
ruby -ane 'puts $F.uniq * ","'

Answer 4

I had similar problem today and I solved it by using echo,tr,sed. Code for your example is as below -
echo -n "18 18 21 21 21 21 18 21" | tr -s ' ' '\n' | sort -u | tr '\n' ',' | sed 's/,$//'
output- 18,21

Answer 5

echo '18 18 21 21 21 21 18 21' |
mawk 'END { print "\n" } !___[$_]-- && $!NF= NF < ++__\
? ","$_ : $_' ORS= RS='[\t- ]+'
18,21
or if u don't mind chaining up 2 awks :
mawk '!__[$_]--' ORS=',' RS='[\t- ]+' | mawk 7 RS=',$'
18,21