Parsing data from delimited blocks - perl

Question

I have a log file content many blocks /begin CHECK ... /end CHECK like below:
/begin CHECK
Var_AAA
"Description AAA"
DATATYPE UBYTE
Max_Value 255.
ADDRESS 0xFF0011
/end CHECK
/begin CHECK
Var_BBB
"Description BBB"
DATATYPE UBYTE
Max_Value 255.
ADDRESS 0xFF0022
/end CHECK
...
I want to extract the variable name and its address, then write to a new file like this
Name Address
Var_AAA => 0xFF0011
Var_BBB => 0xFF0022
I am just thinking about the ($start, $keyword, $end) to check for each block and extract data after keyword only
#!/usr/bin/perl
use strict;
use warnings;
my $input = 'input.log';
my $output = 'output.out';
my ( $start, $keyword, $end ) = ( '^\/begin CHECK\n\n', 'ADDRESS ', '\/end CHECK' );
my #block;
# open input file for reading
open( my $in, '<', $input ) or die "Cannot open file '$input' for reading: $!";
# open destination file for writing
open( my $out, '>', $output ) or die "Cannot open file '$output' for writing: $!";
print( "copying variable name and it's address from $input to $output \n" );
while ( $in ) { #For each line of input
if ( /$start/i .. /$end/i ) { #Block matching
push #block, $_;
}
if ( /$end/i ) {
for ( #block ) {
if ( /\s+ $keyword/ ) {
print $out join( '', #block );
last;
}
}
#block = ();
}
close $in or die "Cannot close file '$input': $!";
}
close $out or die "Cannot close file '$output': $!";
But I got nothing after execution. Can anyone suggest me with sample idea?

Answer 1

Most everything looks good but it's your start regex that's causing the first problem:
'^\/begin CHECK\n\n'
You are reading lines from the file but then looking for two newlines in a row. That's not going to ever match because a line ends with exactly one newline (unless you change $/, but that's a different topic). If you want to match the send of a line, you can use the $ (or \z) anchor:
'^\/begin CHECK$'
Here's the program I pared down. You can adjust it to do all the rest of the stuff that you need to do:
use v5.10;
use strict;
use warnings;
use Data::Dumper;
my ($start, $keyword, $end) = (qr{^/begin CHECK$}, qr(^ADDRESS ), qr(^/end CHECK));
while (<DATA>) #For each line of input
{
state #block;
chomp;
if (/$start/i .. /$end/i) #Block matching
{
push #block, $_ unless /^\s*$/;
}
if( /$end/i )
{
print Dumper( \#block );
#block = ();
}
}
After that, you're not reading the data. You need to put the filehandle inside <> (the line input operator):
while ( <$in> )
The file handles will close themselves at the end of the program automatically. If you want to close them yourself that's fine but don't do that until you are done. Don't close $in until the while is finished.

Answer 2

using the command prompt in windows. In MacOS or Unix will follow the same logic you can do:
perl -wpe "$/='/end CHECK';s/^.*?(Var_\S+).*?(ADDRESS \S+).*$/$1 => $2\n/s" "your_file.txt">"new.txt
first we set the endLine character to $/ = "/end CHECK".
we then pick only the first Var_ and the first ADDRESS. while deleting everything else in single line mode ie Dot Matches line breaks \n. s/^.*?(Var_\S+).*?(ADDRESS \S+).*$/$1 => $2\n/s.
We then write the results into a new file. ie >newfile.
Ensure to use -w -p -e where -e is for executing the code, -p is for printing and -w is for warnings:
In this code, I did not write the values to a new file ie, did not include the >newfile.txt prt so that you may be able to see the result. If you do include the part, just open the newfile.txt and everything will be printed there

Answer 3

Here are some of the issues with your code
You have while ($in) instead of while ( <$in> ), so your program never reads from the input file
You close your input file handle inside the while read loop, so you can only ever read one record
Your $start regex pattern is '^\/begin CHECK\n\n'. The single quotes make your program search for backslash n backslash n instead of newline newline
Your test if (/\s+ $keyword/) looks for multiple space characters of any sort, followed by a space, followed by ADDRESS—the contents of $keyword. There are no occurrences of ADDRESS preceded by whitespace anywhere in your data
You have also written far too much without testing anything. You should start by writing your read loop on its own and make sure that the data is coming in correctly before proceeding by adding two or three lines of code at a time between tests. Writing 90% of the functionality before testing is a very bad approach.
In future, to help you address problems like this, I would point you to the excellent resources linked on the Stack Overflow Perl tag information page
The only slightly obscure thing here is that the range operator /$start/i .. /$end/i returns a useful value; I have copied it into $status. The first time the operator matches, the result will be 1; the second time 2 etc. The last time is different because it is a string that uses engineering notation like 9E0, so it still evaluates to the correct count but you can check for the last match using /E/. I've used == 1 and /E/ to avoid pushing the begin and end lines onto #block
I don't think there's anything else overly complex here that you can't find described in the Perl language reference
use strict;
use warnings;
use autodie; # Handle bad IO status automatically
use List::Util 'max';
my ($input, $output) = qw/ input.log output.txt /;
open my $in_fh, '<', $input;
my ( #block, #vars );
while ( <$in_fh> ) {
my $status = m{^/begin CHECK}i .. m{^/end CHECK}i;
if ( $status =~ /E/ ) { # End line
#block = grep /\S/, #block;
chomp #block;
my $var = $block[0];
my $addr;
for ( #block ) {
if ( /^ADDRESS\s+(0x\w+)/ ) {
$addr = $1;
last;
}
}
push #vars, [ $var, $addr ];
#block = ();
}
elsif ( $status ) {
push #block, $_ unless $status == 1;
}
}
# Format and generate the output
open my $out_fh, '>', $output;
my $w = max map { length $_->[0] } #vars;
printf $out_fh "%-*s => %s\n", $w, #$_ for [qw/ Name Address / ], #vars;
close $out_fh;
output
Name => Address
Var_AAA => 0xFF0011
Var_BBB => 0xFF0022
Update
For what it's worth, I would have written something like this. It produces the same output as above
use strict;
use warnings;
use autodie; # Handle bad IO status automatically
use List::Util 'max';
my ($input, $output) = qw/ input.log output.txt /;
my $data = do {
open my $in_fh, '<', $input;
local $/;
<$in_fh>;
};
my #vars;
while ( $data =~ m{^/begin CHECK$(.+?)^/end CHECK$}gms ) {
my $block = $1;
next unless $block =~ m{(\w+).+?ADDRESS\s+(0x\w+)}ms;
push #vars, [ $1, $2 ];
}
open my $out_fh, '>', $output;
my $w = max map { length $_->[0] } #vars;
printf $out_fh "%-*s => %s\n", $w, #$_ for [qw/ Name Address / ], #vars;
close $out_fh;