remove lines from file ending with / char

[msx]

Hi Monks, Hoping someone could enlighen me. I have a file containing lines which consists of filenames and directories. I need to strip all empty directory entries in this file. Below is an example of the file contents: Filname is filelist.txt

:BOF
/peter/john/lilly.pdf
/john/peter/jill.pdf
/jimmy/gill/
/wilbur/yankee/
/lola/riley/jibmo.txt
/rolly/polly/quebe.lst
:EOF

I'm trying in vain to remove all lines ending with the / char. Note there could be any number of lines ending with / and they could be anywhere in the file. Any wisdom will be greatly appreciated.

[swampyankee]

The first place to look is in perlfaq5. Anyhow, there are several ways to do it. The easiest is to

(((pseudocode)))
open, for reading, the file you want to change (input file)
open, for writing, a temp file.

until EOF {
     read a record from the input file.
          if you want to keep it
               write it to the temp file
}

replace the input file with the temp file

If the file is small (ymmv), you can do something like this:

(((pseudocode)))

open input file for reading

slurp ENTIRE input file into array (@contents = <INFILE>)

close the input file

use grep to remove lines with trailing slashes from the array (@conten
+ts = grep { !/\/$/ } @contents)

open the input file for writing (this act will destroy the input file!
+)

write array into the file

A couple of suggestions: first, test from a copy of the file you plan on doing this to. Second, study perlre and grep for hints. Third, beware of trailing white space.

[choroba]

Please, give us more information: Is it possible for a directory to be mentioned once with no content, but later with a file, like this?

/foo/bar/
/foo/bar/baz

Or only the "empty" directories can be listed without files. If so, you can just use

   grep { 0 != index reverse($_), '/' } @files

or

   grep !m{/$}, @files

[BillKSmith]

This task can be done with a one-liner. Refer to the runtime options /e, /i, and /p in perldoc perlrun.

[philiprbrenan]

Just supposing the problem was to find the smallest set of directories whose removal would not delete any files (the folded comments make it more difficult to read):

use feature ":5.14";
use warnings FATAL => qw(all);
use strict;
use Data::Dump qw(dump pp);

sub findEmptyDirectories(&@)
 {package FindEmptyDirectories;
  my $R = shift;                          # Processing routine from us
+er followed by file/directory names

  my $d = {};                             # Simulated directory tree
   {for(@_)                               # Load directory tree with e
+ach file
     {if (my @c = /([^\/]+)/g)            # Get path components from f
+ile name
       {my $f = pop @c unless /\/\Z/;     # Get file name unless direc
+tory
        my $c = $d;                       # Start of descent through d
+irectory tree
        $c = $c->{dir}{$_} //= {} for @c; # Descend along each path co
+mponent adding new nodes as necessary
        $c->{file}{$f}++ if defined $f;   # Count files at this level 
+with this name
       }
     }
   }

  sub R($)                                # Descend tree to count file
+s in each directory 
   {my ($d) = @_;                         # Position in directory tree
    $d->{files} += &R($d->{dir}{$_}) for keys %{$d->{dir}}; # Count fi
+les below this level
    $d->{files} += keys %{$d->{file}};    # plus files at this level
   }
  R($d);                                  # Count files at each level

  sub r($$$)                              # Descend tree in in-order i
+n order to perform user action on empty directories 
   {my ($R, $d, $p) = @_;                 # User action, directory sub
+ tree, path to sub tree
  
    return $R->(@$p) unless $d->{files};  # No files below this direct
+ory so perform user action and return

    &r($R, $d->{dir}{$_}, [@$p, $_]) for sort keys %{$d->{dir}}; # Els
+e process lower levels in directory tree 
   }
  r($R, $d, []);                          # Call user action on direct
+ories with no files
 }

findEmptyDirectories {say "/", join("/", @_), "/"} split /\n/, << 'END
+';
/peter/john/lilly.pdf
/john/peter/jill.pdf
/jimmy/gill/a.txt
/jimmy/gill/johnny/
/wilbur/yankee/
/wilbur/yankee/aaa/bbb/nnn/
/wilbur/yankee/aaa/ccc/nnn/
/wilbur/yankee/aaa/ccc/aaa.txt
/lola/riley/jibmo/txt
/lola/riley/jibmo.txt
/rolly/polly/quebe.lst
END

Produces:

/jimmy/gill/johnny/
/wilbur/yankee/aaa/bbb/
/wilbur/yankee/aaa/ccc/nnn/