There was a similar question posted here a few days ago. It boils down to this: Perl cannot know where, within a file, "Confname" occurs, without first skimming through the file to find "Confname". Similarly, it cannot know where line 34 begins, how many characters are in the 3000th line, or any other internal file information, without first skimming through the file to find out.

This fact is not constrained to Perl. It is true of any programming language, given the same scenario: find "Confname" in a file. It's a fact that is not even constrained to just the world of computers.

Quick, find me the 23rd occurrence of the word "Radley" in To Kill a Mockingbird, but do so without opening the book, and without skimming your way through the book. It's not possible.

A good piece of advice when thinking through a problem is to think of how you might do it yourself, without a computer. And then program the computer to do what you would have done. If it were you, yourself, looking for "Confname" in a document, you would physically skim through the document looking for that word.

That said, your existing snippet is probably the simplest way to do just that. If you find that the simple solution fails to meet your needs in some way, think of how you would improve the situation in the non-computer world. Maybe after you found "Confname" for the first time, in the non-computer world, you would put a bookmark on the proper page of the document. In the computer world, you might maintain an index file to accomplish the same efficiency.

Hi

Thanks a lot for your clear answer. It is very much useful for me and others who wants to do this task.

"Keep pouring your ideas"

You can also:

while ($fh) {
  /^Confname/ or next;
  work_with $_;
  last;   ## if you want to process only the first ocurrence of such l
+ine
}
[download]

If the lines in the file are not very long, it may fit in memory:

my @confnames = grep /^Confname/, <$fh>;
## if you want to process all lines:
work_with $_ for @confnames;
## else:
work_with $confnames[0];
[download]

Update: Fixed a missing <>. Thanks to Tanktalus!

I'd refactor that slightly to:

while ( <FH> ) {
    next if $_ !~ /^Confname/;
    ...
}
[download]

Or simply, if we're refactoring,

while ( <FH> ) {
    next unless /^Confname/;
    ...
}
[download]

I've always found unless clearer than a negated if for short single-condition conditionals like that. I do accept that that's personal preference, though .. :)

-- Tanalis
#include www.liquidfusion.org.uk