How to reorder a text file

hennesse has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: How to reorder a text file by tybalt89 (Monsignor) on Jun 11, 2018 at 22:06 UTC
`#!/usr/bin/perl # https://perlmonks.org/?node_id=1216436 use strict; use warnings; print reverse split /^(?=Title)/m, do{ local $/; <DATA>}; __DATA__ Title 1 Line of text A Line of text B Title 2 Line of text C Title 3 Title 4 Line of text D` [download] Outputs: `Title 4 Line of text D Title 3 Title 2 Line of text C Title 1 Line of text A Line of text B` [download] Which I think is more valid than your Desired Output	[reply] [d/l] [select]
Re^2: How to reorder a text file by hennesse (Beadle) on Jun 11, 2018 at 23:20 UTC
Works perfectly - Thank you!	[reply]
Re: How to reorder a text file by Cristoforo (Curate) on Jun 11, 2018 at 23:43 UTC
Taken from Re: Split and print hash based on regex `#!/usr/bin/perl use strict; use warnings; open my $fh, '<', \<<EOF; Title 1 Line of text A Line of text B Title 2 Line of text C Title 3 Title 4 Line of text D EOF { local $/; print reverse <$fh> =~ /(^Title(?:(?!^Title).)*)/msg; }` [download]	[reply] [d/l]
Re: How to reorder a text file by kcott (Archbishop) on Jun 12, 2018 at 07:40 UTC
G'day Dave, Take a look at the recent "Sort text by Chapter names" thread. A number of the solutions presented there are easily adapted to your requirements. — Ken	[reply]
Re: How to reorder a text file by tybalt89 (Monsignor) on Jun 12, 2018 at 19:08 UTC
Just for fun, here's a solution that doesn't keep blocks of text in memory, and therefor may work for larger files. Besides, how often does one get to use unshift, and set $/ to a reference :) Notice also that you can seek around in DATA. `#!/usr/bin/perl # https://perlmonks.org/?node_id=1216436 use strict; use warnings; my @sections; while( <DATA> ) { if( /^Title/ ) { unshift @sections, [ tell(DATA) - length, length ]; } elsif( @sections ) { $sections[0][1] += length; } } for ( @sections ) { seek DATA, $_->[0], 0 or die; local $/ = \$_->[1]; print scalar <DATA>; } __DATA__ Title 1 Line of text A Line of text B Title 2 Line of text C Title 3 Title 4 Line of text D` [download]	[reply] [d/l]
Re: How to reorder a text file by johngg (Canon) on Jun 12, 2018 at 17:34 UTC
An approach reading a line at a time rather than slurping the whole file. `use strict; use warnings; open my $inFH, q{<}, \ <<__EOD__ or die $!; Title 1 Line of text A Line of text B Title 2 Line of text C Title 3 Title 4 Line of text D __EOD__ my @groups; while ( <$inFH> ) { if ( m{^Title} ) { unshift @groups, [ $_ ]; } else { push @{ $groups[ 0 ] }, $_; } } while ( my $group = shift @groups ) { print @{ $group }; }` [download] The output. `Title 4 Line of text D Title 3 Title 2 Line of text C Title 1 Line of text A Line of text B` [download] I hope this is of interest. Cheers, JohnGG	[reply] [d/l] [select]
Re^2: How to reorder a text file by AnomalousMonk (Archbishop) on Jun 12, 2018 at 19:32 UTC
... rather than slurping the whole file. But won't the `while ( <$inFH> ) { ... }` loop end up reading the entire file into the `@groups` array (actually an array-of-arrays or AoA) before it finishes? This is not what I think of when I think of "line-by-line" processing of a file. (Actually, it looks a lot like slurping! :) Give a man a fish: `<%-{-{-{-<`	[reply] [d/l] [select]
Re: How to reorder a text file by tybalt89 (Monsignor) on Jun 13, 2018 at 03:26 UTC
I wanted to do two things - read a section at a time, and write the output file from back to front. This allows a single pass through the input file. Since the sections don't have an obvious ending, reading into the next section is required, but that's easy to fix when it happens :) #!/usr/bin/perl # https://perlmonks.org/?node_id=1216436 use strict; use warnings; my $inputfile = 'd.1216436.in'; my $outputfile = 'd.1216436.out'; my $pointer = -s $inputfile; # output will be same size as i +nput open my $in, '<', $inputfile or die "$! opening $inputfile"; open my $out, '>', $outputfile or die "$! opening $outputfile"; local $/ = "\nTitle"; # too far, but we will fix it while( <$in> ) # read whole section { s/\n\KTitle\z// and seek $in, -5, 1; # if read too far, fix & backup seek $out, $pointer -= length, 0; # backup output pointer print $out $_; # put in proper place } close $out; [download] Now if I can only stop giggling :)	[reply] [d/l]
Re: How to reorder a text file by monsenhor (Novice) on Jun 13, 2018 at 01:48 UTC
`#!/usr/bin/env perl use strict; use warnings; my $txt = " Title 1 Line of text A Line of text B Title 2 Line of text C Title 3 Title 4 Line of text D "; my @reverse; my @normal = split /(Title \d)/, $txt; foreach my $n (0 .. @normal){ if ($n%2){ push @reverse, $normal[$n].$normal[$n+1]; } } chomp @reverse; @reverse = reverse @reverse; map { print $_."\n" } @reverse;` [download]	[reply] [d/l]
Re^2: How to reorder a text file by AnomalousMonk (Archbishop) on Jun 13, 2018 at 12:00 UTC
In the `foreach my $n (0 .. @normal){ if ($n%2){ push @reverse, $normal[$n].$normal[$n+1]; } }` [download] loop, aren't you indexing beyond the end of the `@normal` array? Won't this produce `"Use of uninitialized value..."` warnings if warnings are enabled? Shouldn't the loop range be `for my $n (0 .. $#normal-1) { ... }` instead? (And BTW: Aren't you losing an opportunity to avoid using a `foreach`-loop by not re-writing the loop as a ~~`map`~~ \| void-context map statement? :) Give a man a fish: `<%-{-{-{-<`	[reply] [d/l] [select]
Re: How to reorder a text file by rminner (Chaplain) on Jun 13, 2018 at 23:57 UTC
There are already plenty good answers. Adding another answer for the fun of it. This answer uses the CPAN Module File::ReadBackwards. It reads the file backwards and prints out the results every time a block is complete (Title line found). Memory usage should therefore be low. Exit if the expected "Title" line is not found for $MAX_LENGTH chars. Otherwise large input files without "Title" lines could still cause massive memory usage. Caveat: the length is only updated every time a "line" has been read. If there is a huge input line (e.g. several gigabytes without newline), this programm will still crash and burn :). use strict; use warnings; use File::ReadBackwards; my $MAX_LENGTH = 1_000_000; my $infile = shift @ARGV; die "'$infile' is not a file!" unless (-f $infile); my $bw = File::ReadBackwards->new( $infile ) or die "can't read '$infile' $!"; my (@block, $length, $line); while( defined( $line = $bw->readline ) ) { unshift @block, $line; $length += length $line; if ($line =~ m/^Title\b/) { print @block; @block = (); $length = 0; } if ($length >= $MAX_LENGTH) { die "$length chars read, without finding a 'Title' line\n". "Input file '$infile' probably has the wrong file format." +; } } if (@block) { print STDERR scalar(@block) ." lines not printed out yet!"; } [download]	[reply] [d/l]
Re: How to reorder a text file by Anonymous Monk on Jun 11, 2018 at 21:59 UTC
`Title 3 Line of text D Title 3` [download] what?	[reply] [d/l]
Re^2: How to reorder a text file by hennesse (Beadle) on Jun 11, 2018 at 22:06 UTC
Dang - Fried Brain Syndrome. How's this? Input file: `Title 1 Line of text A Line of text B Title 2 Line of text C Title 3 Title 4 Line of text D` [download] Desired Output: `Title 4 Line of text D Title 3 Title 2 Line of text C Title 1 Line of text A Line of text B` [download]	[reply] [d/l] [select]
Re^3: How to reorder a text file by Anonymous Monk on Jun 11, 2018 at 22:12 UTC
`$ tac file \| perl -ne'push@a,$_;if(/^Title\b/){print reverse@a;@a=()}'` [download]	[reply] [d/l]


No such thing as a small change
	PerlMonks