Perl Hash to CSV - arrays

Question

I have a data file for alarms occuring during the day. The format looks like
2014/04/27-23:42:22.177742- Alarm1
2014/04/27-23:42:22.177744- Alarm2
2014/04/27-23:42:22.177747- Alarm3
2014/04/27-23:42:22.177749- Alarm1
Now I cant guess when any alarm would appear. it depends on the system. What I have done is to insert the data of alarm (e.g Alarm1) into a 2D hash. I take a chunk of 5 minutes everytime and look for alarm that appeared during the 5 minutes. I would add the value into hash, everytime I find a new alarm. In case of a repetition (like Alarm1 above) I would simply add 1 to the value. So in the end it would give me a hash which contains the alarm name and the times it appeared in 5 minutes.
Next I would start processing the next 5 minutes.
I am processing it for the whole day, so it is possible that 1 alarm can start appearing 10 in the morning, so It would be a new entry in to the hash. Now ehen I try to print the values eventually to a CSV, its a mess. Totally makes no sense. What I expect is a csv which should look like
Name,00:00,00:05,00:10,
Alarm1,2,5,2,7,
Alarm2,4,7,3,6
Alarm3,6,1,6,3
...
My code is:
use Time::Local;
use POSIX 'strftime';
use Data::Dumper;
my %outputHash= ();
$curr = timelocal(0, 0, 0, (split /\//, $ARGV[0])[1], (split /\//, $ARGV[0])[0]-1, (split /\//, $ARGV[0])[-1]);
$currentTime = strftime "%Y/%m/%d-%H:%M:%S", localtime($curr);
for ($count = 1; $count <= 288; $count++) { #there are 288 '5 minutes' in a day.
$curr += 300;
$nextTime = strftime "%Y/%m/%d-%H:%M:%S", localtime($curr);
$cmd = "awk '\$0>=from&&\$0<=to' from=\"$currentTime\" to=\"$nextTime\" Output.txt";
my $dataChunk = qx($cmd);
my #lines = split /[\n]+/, $dataChunk;
foreach my $line (#lines) {
chomp;
$timeStamp1 = substr($line,21,6);
#print "\n$timeStamp1\n$error\n";
if ($timeStamp1 != $timeStamp2){
$outputHash{$error}{$count} = $outputHash{$error}{$count} + 1;
}
$ind = index($line,'- ') + 2;
$len = length($line) - $ind;
$error = substr($line,$ind, $len);
$timeStamp2 = $timeStamp1;
}
$currentTime = $nextTime;
# if ($count>3){$count=300;}
}
`>/tmp/report.txt`;
open (MYFILE, '>>/tmp/report.txt');
my #outputArray = ();
my $flag =1;
foreach my $error (sort keys %outputHash)
{
print MYFILE "$error,";
#$outputArray[$flag][0] = $error;
for ($count=1,$count <= 288, $count++)
{
print MYFILE "$outputHash{$error}{$count},";
#$outputArray[$flag][$count] = int($outputHash{$error}{$count});
}
$flag += 1;print MYFILE "\n";
}
close (MYFILE);
#print Dumper(\#outputArray);
exit;
A simplified display of my has looks like this. The reason its haphazard is because Alarm 1 occurred only in '2nd' 5 minutes interval, Alarm 2 occured in 1st only, Alarm 3 occured in 4 consecutive 5 minutes intervals we monitored.
'Alarm1{
'2' => '5'
},
'Alarm2{
'1' => '1'
},
'Alarm3
'4' => '1',
'1' => '2',
'3' => '1',
'2' => '1'
},

Answer 1

Try this out, its best if you use a module that is meant for dealing with CSV.
I chose Class::CSV because it is is simple to use.
#!/usr/bin/perl
use strict;
use warnings;
use Class::CSV;
my %hash = (
'Alarm1' => {'2' => '5', },
'Alarm2' => {'1' => '1', },
'Alarm3' => {
'4' => '1',
'1' => '2',
'3' => '1',
'2' => '1'
},
);
my #fields = qw/AlarmNo 00:00:00 00:05:00 00:10:00 00:15:00/;
my $csv = Class::CSV->new( fields => \#fields );
#make the hash into a suitable array
my #array;
my #keys = keys %hash;
for my $i (0 .. $#keys){
push #{ $array[$i] }, $keys[$i];
for my $inter (1 .. 4){
my $val = '';
if(exists $hash{$keys[$i]}->{$inter}){
$val = $hash{$keys[$i]}->{$inter};
}
push #{ $array[$i] }, $val;
}
}
$csv->add_line($_) for(#array);
print join(',', #fields), "\n"; #Just to make it tidy on the commandline
$csv->print();
So you would use print MYFILE $csv->string to get it into your file.
Edit:
If you can't install Class::CSV check out Text::CSV which may be installed by default.
You can also join the arrays by a comma like so
for(#array){
print join(',', #{$_});
}