In a list of file paths in an RTF file, count and sort output based on number of occurences of each file name

Using Raku (formerly known as Perl_6)

~$ raku -ne 'BEGIN my %seen; 
             if .chars && /\.pdf/ { $_.subst(/ <?after \.pdf> \: \d+ $ /).IO.basename andthen %seen{$_}++ }; 
             END .say for %seen.sort: +*.key.match(/^ \d+ <?before \.pdf>/);'   file

OR (more simply):

~$ raku -ne 'BEGIN my %seen; 
             if .chars && s/ <?after \.pdf> \: \d+ $ // { %seen{$_.IO.basename}++ }; 
             END .say for %seen.sort: +*.key.match(/^ \d+ <?before \.pdf>/);'   file

OR (even more simply):

~$ raku -ne 'BEGIN my %seen; 
             if .chars && s/ <?after \.pdf> \: \d+ $ // { %seen{$_.IO.basename}++ }; 
             END .say for %seen.sort: +*.key.IO.extension: "";'   file

Raku is a programming language in the Perl family. Using Raku's awk-like -ne non-autoprinting commandline flags, you can obtain a sorted hash of key/value pairs where the value per PDF file name equals the count of the number of times that filename was seen. Output is sorted numerically according to filename (as a number). Using .say in the END block will give you paired "key => value" output:

Sample Input:

Category1:
./Folder1/Folder2/1.pdf:18
./Folder3/2.pdf:18
./Folder5/4.pdf:10

Category2:
./Folder3/2.pdf:18
./Folder5/4.pdf:10

Category3:
./Folder1/Folder2/1.pdf:18
./Folder5/4.pdf:10

Category4:
./Folder6/7.pdf:10
./Folder5/4.pdf:10
./Folder3/2.pdf:18

Sample Output:

1.pdf => 2
2.pdf => 3
4.pdf => 4
7.pdf => 1

If you need paired paired "key : value" output, change the END block of your code to:

 END put($_.key, " : ", $_.value) for %hash.sort: +*.key.IO.extension: "";'

If you need to eliminate pairs with .value < 2, further change the END block of your code to:

 END put($_.key, " : ", $_.value if .value > 2 ) for %hash.sort: +*.key.IO.extension: "";'

Finally, if you prefer code written in a "chained" method/function call -style, the code below gives the same (desired) code as that above:

~$ raku -e 'my  %seen = lines.grep( *.chars > 0 && / \.pdf /)  \ 
                        .map( *.subst(/ <?after \.pdf> \: \d+ $ / ).IO.basename).Bag;  \ 
            for %seen.sort( +*.key.IO.extension: "") { 
                put $_.key ~" : "~ $_.value if .value > 2 };'   file

https://raku.org

With perl:

$ perl -lne '$c{$1}++ if m{([^/]*\.pdf):\d+$};
             END {
               print "$_ : $c{$_}" for
                 sort {$c{$b} <=> $c{$a}} grep {$c{$_} > 2} keys %c
             }' your-file
4.pdf : 4
2.pdf : 3

Extract that file to be a TEXT file, then do

$ sed < TEXT.txt -nre  's,.*/([^:/]+).*,\1,p' | sort | uniq -c
      2 1.pdf
      3 2.pdf
      4 4.pdf
      1 7.pdf

$ sed < TEXT.txt -nre  's,.*/([^:/]+).*,\1,p' | sort | uniq -c | grep -vE '\s+[12] '
      3 2.pdf
      4 4.pdf

I would probably use grep here.

$ grep -Eo '[^/]+\.pdf' ./input.txt |
                               sort |
                            uniq -c | awk '$1>2 { print $2" : "$1 }'
2.pdf : 3
4.pdf : 4

Steps:

• grep -Eo '[^/]+\.pdf' selects and prints just the PDF filenames.
• sort | uniq -c counts the distinct filenames.
• awk '$1>2 { print $1" : "$2 }' selects all lines with PDF names that occur at least 3 times, and formats them in the desired way.

A couple related problems:

• What if spaces are allowed in the filenames?
  We can handle that by making the AWK expression a little uglier:

  grep -Eo '[^/]+\.pdf' ./inp.txt |
                             sort |
                          uniq -c | awk '$1>2 { n=$1; $1=""; print n" : "$0 }'
  

• What if we want to count ./folder1/x.pdf and ./folder2/x.pdf as different?
  The best thing is probably to just edit the grep expression so it prints the full path, then report the files as unique up to paths:

  grep -Eo '^.*+\.pdf' ./inp.txt |
                             sort |
                          uniq -c | awk '$1>2 { print $2" : "$1 }'