I am reading a file and initializing the hash with respective segment. And i have used following loop,
I am wondering if there is any smart way of doing this task, i mean better than what I have done.
following is my code,
[code]
while (my $line = <FILEREAD>)
{
$line =~ s/\r\n//; #Subsitute binary character \cM\cJ
chomp($line);
if ($line)
{
if ($line eq "[RESOURCE]"){$RFlag = "WHITE"};
if ($line eq "[WNT]"){$WFlag = "WHITE";$RFlag = "RED"};
if ($line eq "[UNX]"){$UFlag = "WHITE";$WFlag = "RED"};
if ($RFlag eq "WHITE")
{
if ($line =~ /^Id.*?$/)
{
$key = $line;
}
elsif ($line =~ /^Source.*?$/)
{
$$HR_resource{$key} = "$line";
}

}
if ($WFlag)
{
if ($WFlag eq "WHITE")
{
if ($line =~ /^Id.*?$/)
{
$key = $line;
}
elsif ($line =~ /^Source.*?$/)
{
$$HR_Window{$key} = "$line";
}

}
}
if ($UFlag)
{
if ($UFlag eq "WHITE")
{
if ($line =~ /^Id.*?$/)
{
$key = $line;
}
elsif ($line =~ /^Source.*?$/)
{
$$HR_Linux{$key} = "$line";
}

}
}
}
[\code]

Any suggestion welcome,

Cheers

Just judging by the code and not knowing anything else, it looks like it could be polished up a bit (and the syntax errors have to be removed for it to even work).

Why are you using soft references for the hashes? That should alomst certainly be switched to a regular hash instead of a soft reference to a hash. You could probably use one hash (as in a hash of hashes) instead of three hashes but there is no way to know without further information.

This post has been edited by KevinADC: 18 Jun, 2009 - 06:49 PM

Could you please guide me how to get rid of multiple hashs
and actully i am looking for some other way so that i don;t have to use those FLAG's
any suggestion
let me know

Cheers

Without further information I don't want to try and devine what your code is doing. Post some sample data and explain what you need to do with the data.

Let's take a look, hmm?


These two lines are redundant. chomp removes the trailing (OS-dependent) newline from a value if there is a trailing newline. In theory, the chomp alone should be sufficient, but, in practice, it doesn't work so well if the file's line endings are different from those on the OS where the code is run, so I'd suggest using "$line =~ tr/\r\n//d" on its own, as this will remove any and all \r or \n characters from the line, whether they appear individually or paired (and tr/// runs faster than a full-blown regex).


This looks to me like you're trying to implement a simple state machine using flags to indicate whether you're in a "RESOURCE", "WNT", or "UNX" section, but it's unclear why you're using two flags instead of just one. I suspect that

would be an easier and cleaner way to accomplish this, plus it will automatically catch any lines containing "[ALL CAPS TEXT]" if new section types are added later, although you may need to alter the regex to allow a broader range of characters than just uppercase letters and spaces.

This method is also less fragile in the case of files with out-of-order sections. The rest of your code seems to assume that only one of $RFlag, $WFlag, or $UFlag will be "WHITE" at any given time, but an input file with the [UNX] section first, then [WNT], then [RESOURCE] would set all three to "WHITE", violating this assumption.

The remaining code appears to be identical for all three sections, aside from which hash the "Source" values are inserted into, so they can be stripped down to:

$results{RESOURCE}, $results{WNT}, and $results{UNX} will then correspond to your \%HR_resource, \%HR_Window, and \%HR_Linux, respectively.

Also note that I removed the double quotes around "$line" in the assignment to the hash and the trailing ".*?$" from the regexes, as they were superfluous.

Putting this all together, we get (syntactically valid, but completely untested):

which should behave identically to the code you posted, within the assumed constraints that input files must contain a [RESOURCE] section followed by a [WNT] section and then a [UNX] section and that you don't care what happens with data found in any section(s) other than those three.

Your solution looks really impressive, but really important thing I am unsure is about how to extract respective hash from following:-


Any suggestion, welcome

Cheers

As I said in the earlier message,

so you should (again, untested, and I'm not sure of the context your original code came from, so I can't be 100% certain) be able to just replace:

$$HR_resource{...} with $$$results{RESOURCE}{...}
%$HR_resource with %{$$results{RESOURCE}}

throughout the rest of your code (along with the corresponding HR_Window -> results{WNT} and HR_Linux -> results{LNX} constructs, of course).