Re: How to match the each line of of users for a perticular time with predeclared group.

Couple of points.

If you want to lookup each users group then one way is to build a lookup table (a hash).

I would print the report after all the data has been parsed building an output hash as you go.

I've used the user names you used in your sample data to keep things simple.

#!/usr/bin/perl 

use strict;
use warnings;
use Data::Dumper;

my %groups;
$groups{'g1'} = ['a'..'g'];
$groups{'g2'} = ['h'..'n'];
$groups{'g3'} = ['o'..'u'];
$groups{'g4'} = ['v'..'z'];

my %lookup;
for my $grp (keys %groups){
  %{$lookup{$grp}} = map {$_ => undef} @{$groups{$grp}};
}

my %output;
while(my $line=<DATA>){
  chomp($line);
  my $usr=substr($line,8,2);
  my $time=substr($line,0,4); 
  my $username_string = substr($line,11);
  my @usernames= split ',', $username_string;
  for my  $name (@usernames){
    for my $grp (keys %lookup){
      $output{$time}{$grp}++ if exists $lookup{$grp}{$name};
    }
  }
}

for my $time (sort keys %output){
  print "$time\t";
  for my $grp (sort keys %groups){
    my $logged_in = 0;
    $logged_in = $output{$time}{$grp} 
      if exists $output{$time}{$grp};
    print "$logged_in ($grp) users\t";
  }
  print "are logged in\n";
}

__DATA__
1.00 50 10 x,y,z,w,u,e,g,j,k,a
1.10 50 13 x,y,b,a,g,j,k,r,,n,m,s,
1.20 50 05 c,t,g,q,f
1.50 50 08 a,t,y,w,z,x,s,b
2.00 50 14 x,y,u,b,g,f,s,a,i,o,p,c
[download]

output:

---------- Capture Output ----------
> "C:\Perl\bin\perl.exe" monk.pl

1.00    3 (g1)    2 (g2)    1 (g3)    4 (g4)    users logged in
1.10    3 (g1)    4 (g2)    2 (g3)    2 (g4)    users logged in
1.20    3 (g1)    0 (g2)    2 (g3)    0 (g4)    users logged in
1.50    2 (g1)    0 (g2)    2 (g3)    4 (g4)    users logged in
2.00    5 (g1)    1 (g2)    4 (g3)    2 (g4)    users logged in

> Terminated with exit code 0.
[download]

updated: tidied up the output.

Comment on Re: How to match the each line of of users for a perticular time with predeclared group. Select or Download Code

Replies are listed 'Best First'.
Re^2: How to match the each line of of users for a perticular time with predeclared group. by jhourcle (Prior) on Jan 18, 2007 at 18:45 UTC
If the group assignments are mutually exclusive (no person is assigned to more than one group ... which doesn't look to be the case, as 'apatnayi' is in groups 2 and 4), rather than loop through each of the groups, it's easier to have the lookup be a mapping of the group assignments: `my %lookup = map { my $grp = $_; map { $_ => $grp } ( @{ $groups{$grp} }; } (keys %groups); # ... and then below ... foreach my $name (@usernames) { $output{$time}{ $lookup{$name} }++; # ... or, if you have users logged in that aren't in a group # $output{$time}{ $lookup{$name} }++ if exists $lookup{$name}; # ... or, to track those users with no groups # $output{$time}{ $lookup{$name} \|\| 'unknown' }++; }` [download] If it's not, then you can have the lookup be to an array of what groups they're in, so you don't have to look through each group every time: `my %lookup = (); foreach my $grp (keys %groups) { push @{$lookup{$_}}, $grp foreach @{$groups{$grp}}; } ... foreach my $name (@usernames) { $output($time){$_}++ foreach @{$lookup{$name}}; }` [download]	[reply] [d/l] [select]

Replies are listed 'Best First'.

Re^2: How to match the each line of of users for a perticular time with predeclared group.
by jhourcle (Prior) on Jan 18, 2007 at 18:45 UTC

If the group assignments are mutually exclusive (no person is assigned to more than one group ... which doesn't look to be the case, as 'apatnayi' is in groups 2 and 4), rather than loop through each of the groups, it's easier to have the lookup be a mapping of the group assignments:

my %lookup = map {
  my $grp = $_;
  map { $_ => $grp } ( @{ $groups{$grp} };
} (keys %groups);

# ... and then below ...

foreach my $name (@usernames) {
    $output{$time}{ $lookup{$name} }++;
    # ... or, if you have users logged in that aren't in a group
    # $output{$time}{ $lookup{$name} }++ if exists  $lookup{$name};
    # ... or, to track those users with no groups
    # $output{$time}{ $lookup{$name} || 'unknown' }++;
}
[download]

If it's not, then you can have the lookup be to an array of what groups they're in, so you don't have to look through each group every time:

my %lookup = ();
foreach my $grp (keys %groups) {
    push @{$lookup{$_}}, $grp foreach @{$groups{$grp}};
}
...
foreach my $name (@usernames) {
    $output($time){$_}++ foreach @{$lookup{$name}};
}
[download]

[reply]
[d/l]
[select]