Hello anupchandu, and welcome to the Monastery!

The obvious problem is that you are splitting the input on whitespace, but individual log file records also contain embedded whitespace. The solution depends on the way in which the data is delimited within the log file. For example:

• If the records are of fixed lengths, see How can I manipulate fixed record length files?
• If the records are tab-separated (and you can be certain there are no tab characters embedded in the data), just split on /\t/.
• If the records are comma-separated, use a module such as Text::CSV.

If this advice fails to address your problem, you will need to detail both the format of your log file and the way in which your current script is “not successful.”

Hope that helps,

In addition to what Athanasius says, you need to add

before your while loop, and

after it.

Unless, of course you're not showing us the totality of your code (I notice that you have use HTML::Template; at the top of your snippet, but you never actually, er, use it).

Also, what is use Win32; supposed to do?

G'day anupchandu,

If the status and dates contain no spaces, a quick and dirty fix might be:

my $status = shift @cells;
my $end = pop @cells;
my $start = pop @cells;
print "<td>$_</td>" for ($status, join(' ', @cells), $start, $end);
[download]

A more robust solution would be to work on the basis that the formats of the status and dates are known: if you match them, whatever is left in the middle is the company name. Here's an example:

# Do this once:
my $status_re = qr{\w+};
my $date_re = qr{\d{2}-\w{3}-\d{4}};
my $log_re = qr{^($status_re)\s+(.*?)\s+($date_re)\s+($date_re)\R$};
...
# Do this in the while loop
print "<td>$_</td>" for $line =~ $log_re;
[download]

As you can see from earlier responses, our ability to provide you with a definitive answer is hampered by the lack of information in your original post. A better question gets better answers: read the guidelines in "How do I post a question effectively?" to see how you could have improved on this.

This, in my experience, is a common error for someone parsing a log for the first time so don't feel bad.  :-) I prefer to parse logs based on predictable components. The more wild the potential format, the more complicated the code gets, but for a relatively simple format like the one you are suggesting, I think it's fairly straightforward (assuming you have a basic understanding of Regular Expressions).

You have to craft your Regular Expression to match the data you are expecting. A technique I have become fond of is the use of an ifstatement, which provides the additional feature of filtering out lines that don't match my preconceived format. I often capture those out to another file for occasional review to see if the parsing routine needs to compensate for previously unknown formats or conditions. I won't do that in this example so we can save space.

C:\Steve\Dev\PerlMonks\P-2013-10-27@0838-Log-Parse>type test1.log
GOOD Acme Toy Company 2010-01-01 2011-12-31
BAD XYZZY 1972-01-01 1972-06-18
UGLY Enron 2001-10-01 2011-09-11

C:\Steve\Dev\PerlMonks\P-2013-10-27@0838-Log-Parse>parselog.pl test1.l
+og
[download]

Status	Company Name	Start Date	End Date
GOOD	Acme Toy Company	2010-01-01	2011-12-31
BAD	XYZZY	1972-01-01	1972-06-18
UGLY	Enron	2001-10-01	2011-09-11