Parsing file and joining content into string

[Mark.Allan]

Hi Monks

I am not asking for anyone to write any code here but I am looking for good advice where to start with my issue.

I have a flat file and what I need to do is read the contents of that file (which is simple enough) but the part I cant work out is I need to join the contents of the file into one complete string between pointers in a file.

Example is an extract from the test file

# 20130115 175816.654330
modify 0;mc_ueid='mc.hq_i_aix_01.10f59802.0';
        mc_modhist=[hq_i_aix_01];
        repeat_count=1;
END

# 20130115 175817.304403
modify 0;mc_ueid='mc.hq_i_aix_01.10d41bfd.2';
        mc_modhist=[hq_i_aix_01];
        repeat_count=14555;
END

# 20130115 175817.615425
modify 0;mc_ueid='mc.hq_i_aix_01.10c99dfe.1';
        mc_modhist=[hq_i_aix_01];
        repeat_count=14605;
END

# 20130115 175818.571722
modify 0;mc_ueid='mc.hq_i_aix_01.10f58d1f.0';
        mc_modhist=[hq_i_aix_01];
        repeat_count=10;
END

I need to start at the # of each entry and finish at the END and remove the carriage returns and join the contents of the data into one complete string keeping the ";" separators. I need to do this for each occurrence of data which falls between # and END. I hope I've explained it ok

Any help would be much appreciated"

[LanX]

> I am not asking for anyone to write any code here but I am looking for good advice

great! =)

> I need to start at the # of each entry and finish at the END

2 possible approaches

1. you change the INPUT_RECORD_SEPARATOR to 2 linebreaks $/="\n\n"

Iterating now with  while ( $chunk = <$INPUT>) will give you multiple lines to process in chuncks.

2. you use the flip-flop operator '..' to read all lines between from start till end pattern

 if ( $line =~ /^# \d+ \d+\.\d+$/   ..   $line =~ /^END$/ ) {
     $chunk.=$line; 
 } else {  
     process($chunk) if $chunk;
     $chunk='':
}

You're free to directly process($line) and to skip the '$chunk'-part completely.

> join the contents of the data into one complete string keeping the ";" separators.

well $chunk is a complete string now, do a regex that substitutes ";\n" with ";"

If you process line-by-line try chomp $line to kill all '\n'

> I need to do this for each occurrence of data which falls between # and END.

You need to get rid of start and end line?

Then substitute them away with patterns given. see s///

Or extend the flip-flop expression and add  and not //

'//' defaults to matching the last successful pattern, in your case first and last line. not excludes them

So I'm curious to see your code and I hope others won't take the fun away! =)

HTH!

[Old_Gray_Bear]

Take a look at the special variable $/ AKA $INPUT_RECORD_SEPARATOR. If you set it to 'END', then your read will return all the data a paragraph at a time. Slice 'n' dice of the paragraph is left as an exercise.

[flexvault]

Welcome Mark.Allan,

If you loop on the file as:

   my $string = "";
   while ( <$IN> )                    ## $IN needs to be open
   {   chomp;
       if ( $_ eq "END" )
       {   $string = join ( " ", split " ", $string );
           # $string =~ s/$other//g;   ## In case you need to remove o
+ther characters
           print $OUT "$string\n";    ## $OUT is open output file
           $string = ""; next;
       }
       $string .= $_;
   }

I know you didn't want code, but it was easier to type code then describe :-)

That special use of 'join' is in the Camel book (3rd edition) on page 154. It removes whitespace before and after the string, plus reduces multiple spaces to 1. I typed this in, so there may be some typos, but it should get you going, and remember to use 'use strict; use warning;' in your code.

Hope it helps.