treat diff file in perl

[Anonymous Monk]

Hi Esteemed Monks,

how treat a json diff in perl?

my @diff = <<EOF
     @@ -3,7 +3,7 @@;
     id : 6109;
     name: abc;
     value: 0;
    - status: ?;
    + status: C;
      time: 2023-01-01 12:00;
      @@ -3,8 +3,8 @@;
     id : 6109;
     name: abc;
    - value: 0;
    - status: ?;
    + value: 2
    + status: CS;
      time: 2023-01-01 12:03;
    @@ -3,7 +3,7 @@;
     id : 6109;
     name: abc;
     value: 0;
    - status: ?;
    + status: C;
      time: 2023-01-01 12:05;
EOF

how print every change sorted by id like:

                 6109 changed on 2023-01-01 12:00, before status is ? 
+after is C.
                 6109 changed on 2023-01-01 12:03, before value is 0 a
+fter is 2, before status is ? after is CS.
                 ......

I've searched on CPAN, but all module about diff are generate or apply diff but not treat it. is there a module or convenient way to solve this? Thanks in advance.

[Corion]

What you have is not specific to JSON.

Most likely you can simply walk through the diff file and output the changes. If the diff file does not have enough context, you have to read the original and then apply the diff to output the changes.

What code have you already written to process the diff file?

[tybalt89]

#!/usr/bin/perl

use strict; # https://perlmonks.org/?node_id=11152408
use warnings;
$SIG{__WARN__} = sub { die $@ };

# NOTE missing ; was added

my $diff = <<EOF; # NOTE $ not @
     @@ -3,7 +3,7 @@;
     id : 6109;
     name: abc;
     value: 0;
    - status: ?;
    + status: C;
      time: 2023-01-01 12:00;
      @@ -3,8 +3,8 @@;
     id : 6109;
     name: abc;
    - value: 0;
    - status: ?;
    + value: 2;
    + status: CS;
      time: 2023-01-01 12:03;
    @@ -3,7 +3,7 @@;
     id : 6109;
     name: abc;
     value: 0;
    - status: ?;
    + status: C;
      time: 2023-01-01 12:05;
EOF

my @ids;
for ( split /(?=^\s*@@)/m, $diff )
  {
# print "----------------------------------------\n$_\n\n";
  my ($id) = /^\s*id : (\d+)/m;
  my ($time) = /^\s*time: (.+);/m;
  my @fields;
  push @fields, "before $1 is $2 after is $3"
    while /^\s*- (\w+): (.+);\n(?=(?:.*\n)*\s*\+ \1: (.*);\n)/gm;
  local $" = ', ';
  push @ids, "$id changed on $time, @fields.\n";
  }

print sort { $a =~ s/ .*//sr <=> $b =~ s/ .*//sr } @ids;

Outputs:

6109 changed on 2023-01-01 12:00, before status is ? after is C.
6109 changed on 2023-01-01 12:03, before value is 0 after is 2, before
+ status is ? after is CS.
6109 changed on 2023-01-01 12:05, before status is ? after is C.

[hv]

In the general case, there is not enough information in a diff file (in isolation) to be able to give such context - because in the general case there is no guarantee that unchanged entries such as the id will appear in the context lines. Because of that there is unlikely to be a general-purpose module available for this.

In your specific case, maybe it is possible - if the JSON is guaranteed to be short enough that the id and timestamp will always appear in the context lines. In that case, a hand-rolled parser for the diff chunks should be easy to construct and use; you may be able to get a leg-up by finding something that already parses diff files (such as the inline "Patch" class in patch, part of PerlPowerTools), but it may well be overkill for your needs.

[Discipulus]

Hello,

the problem sounded intriguing and, for my own amusement I got this, where the output is not exactly what asked for but still human readable:

perl -n0E "map{for(join'|',split/\n/){$r{$2}=$1if/(.*)time: (.*);/}map
+{s/-/BEFORE/g;s/\+/AFTER/g;s/.*id : (\d+);.*name: \w+;/ IN \1 /;s/[;|
+ ]+/ /g}values%r}split/^.*@;/m;say$_.$r{$_}for sort keys%r" diff-exam
+ple.txt

2023-01-01 12:00 IN 6109 value: 0 BEFORE status: ? AFTER status: C
2023-01-01 12:03 IN 6109 BEFORE value: 0 BEFORE status: ? AFTER value:
+ 2 AFTER status: CS
2023-01-01 12:05 IN 6109 value: 0 BEFORE status: ? AFTER status: C

The oneliner became ugly and longish for the text formatting, while a shorter one still has an acceptable output:

perl -n0E "map{for(join'',split/\n/){$r{$2}=$1if/(.*)time:(.*)/}}split
+/.*@;/m;say$_.$/.$r{$_}for sort keys%r"

L*

[bliako]

Just in case you have the JSON data available and not just their diff:

Read in the JSON data and convert them into a Perl data structure. Then use Struct::Diff to compare them and show their differences (assuming you overcome deciphering diff's output).

bw, bliako