how to recognize the difference between file_1 and a b c, other than p
+ositionally.
Which is of no help, because a b c can change, depending on which
fields I want to read!
[download]

Well, I've got to kick that question back to you. How would you know, as a human being, which arguments on the command line are filenames, and which ones are columns within filenames?

You can't teach a computer to do what you yourself cannot compute, or decide how to compute. So first, teach me, and I can help you figure out the code.

For example, what rule can you use to distinguish a, the filename, from a, the column name?

I think that the way you've stated your problem, it's intractable, which is why I keep kicking it back. You'll need to annotate the column names distinctly from the filenames somehow.

-- Randal L. Schwartz, Perl hacker

Ok, Merlyn, so maybe you and your book learned me something:

#!/usr/local/bin/perl5 

#I'll call this snippet exfields.pl as in  "EXtractFIELDS"

#zcat command for gzipped files 

$gzcatcmd="gunzip -c"; # I'll need this eventually

# Input parameters

$RegExp     = "";      # Default no regexp
$MatchFld   = "";      # Default no field match
$LastHdr    = "";      # Last read header line
$DbFSplit = ' ';  # Need this form for split
$DbOFSep = ' ';   # Default output field delimiter 
$VBarDel = 0;     # Flag default not splitting by vertical bar
($progname = $0) =~  s/^.*\/([^\/*]+)$/$1/;  # name of this program 

while (@ARGV) {
    $Param = shift @ARGV;
    if ($Param eq "-s") {
    $RegExp = shift @ARGV;     # Search expression
    } elsif ($Param eq "-f") {
    $MatchFld = shift @ARGV;   # Search field specifier
    } elsif ($Param eq "-e") {
    $FList = shift @ARGV;      # field extract list
   } elsif ($Param =~ /^\-/) {
    &ErrorExit; #I'll define this later; you get the point
    } else {
    push @FileList,$Param;
    }
}
# Flag to substitute for whitespace if the new field delimiter 
#includes whitespace and the old one did not. The assumption is 
#that if the old delimiter included whitespace, we do not need 
#to convert existing whitespace in fields. May be modified by 
# #DFD declaration..

if (($DbOFSep =~ /^ +$/) &&
    ($DbFSplit !~ /^ +$/)){
    $FixSpaces = 'Y';  # can't have Null fields or whitespace
} else {
    $FixSpaces = '';   # can't have Null fields or whitespace
}

@OrigFields = split(',',$FList);
unless (@FileList) {push @FileList,"-";}

    print " @FileList\n";
[download]

</CODE
If you run this:
<CODE>
exfields.pl -e Name,Rank,SerialNumber file.gz
[download]

it will return:

Name,Rank,SerialNumber
[download]

Great. Now what? I've succesfully coded HALF of what I wanted to do! What I CAN'T figure out how to do, is:

exfields.pl -e Name,Rank,SerialNumber file.gz | Who,What,Where another
+file.gz
[download]

And have it return:

print "@FileList_1\n";   #prints "Name" "Rank" SerialNumber"
print "@FileList_2\n";   #prints "Who" "What" "Where"
[download]

How am I doing as far as explaining what I'm trying to do?

Ok. I'm going to try and code it, as an answer to your question, then let me know your comments..