Is this the correct structure?

[Anonymous Monk]

Hello dear Monks,
I have this file:

METHOD    0TM    1TM    2TM    >2TM
method1    3    68    18    10
method2    3    80    10    6
method3    3    87    9    1
method4    5    83    9    3
method5    6    75    15    4
method6    14    77    6    3
method7    12    82    2    3
method8    9    84    3    3
method9    13    68    16    3
method10    5    65    20    9
method11    9    64    20    7
method12    5    68    18    9
method13    6    82    9    2
method14    11    79    8    2
method15    6    69    20    5
method16    5    80    11    3
method17    1    68    17    14

and I want to end up with 4 arrays, which will be the following ones (but created on-the-fly).

@x = qw(method1 method2 method3 method4 method5 method6 method7 method
+8 method9 method10 method11 method12 method13 method14 method15 metho
+d16 method17); 
@y1 = (3, 3, 3, 5, 6, 14, 12, 9, 13, 5, 9, 5, 6, 11, 6, 5, 1); 
@y2 = (68, 80, 87, 83, 75, 77, 82, 84, 68, 65, 64, 68, 82, 79, 69, 80,
+ 68); 
@y3 = (18, 10, 9, 9, 15, 6, 2, 3, 16, 20, 20, 18, 9, 8, 20, 11, 17); 
@y4 = (10, 6, 1, 3, 4, 3, 3, 3, 3, 9, 7, 9, 2, 2, 5, 3, 14);

Is this correct code?

open IN, "$infile" or die $!; 
@lines = <IN>;
close IN;

#get first line which will be the header line and keep only the titles
+ we need (the column names)
$header_line = shift (@lines);
@split_header = split(/\t/, $header_line);
shift (@split_header);

#work with the rest of the input file
for ($i=0; $i<=$#lines; $i++)
{
    $method_line = $_;    #lines with the methods and the results for 
+each one
    @split_method = split(/\t/, $method_line);
    
    $method_name= shift (@split_method);  #get only the name of the me
+thod
    push @methods, $method_name;

    push @AoA, [ @split_method ];    #keep the data
}
unshift @AoA, [ @methods ];

[moritz]

Is this correct code?

No. To find out yourself, you can do something like

use Data::Dumper;
print Dumper \@AoA;

and you'll see that @AoA doesn't have the structure you want.

[2teez]

Hi,
I think really, you can achieve what you wanted, by combining all your have except that I would not favour C type of "for loop".. See a headup below:

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

<DATA>; #remove the heading if not needed

my @data;
my $counter;

while(<DATA>){
  chomp; 
  $counter = 0;
  for (split/\s+/,$_){
     push @{$data[$counter++]},$_;
  }
}

print Dumper \@data;




__DATA__
METHOD        0TM        1TM        2TM        >2TM
method1        3        68        18        10
method2        3        80        10        6
method3        3        87        9        1
method4        5        83        9        3
method5        6        75        15        4
method6        14        77        6        3
method7        12        82        2        3
method8        9        84        3        3
method9        13        68        16        3
method10    5        65        20        9
method11    9        64        20        7
method12    5        68        18        9
method13    6        82        9        2
method14    11        79        8        2
method15    6        69        20        5
method16    5        80        11        3
method17    1        68        17        14

You get an ARRAY of ARRAY, then you can print out want you want.
for more info. Please check perldsc.
Update:
The code above produces ...
Read more... (3 kB)

[Anonymous Monk]

Thank you all!
Here is where I am at right now:


#!/usr/bin/perl
use Chart::Gnuplot; 

#one must specify FILENAME to get data from (tab-delimited), a name fo
+r the output file, a type for output file (e.g. pdf, png, eps) and a 
+plot title
($infile, $title, $type, $plot_title) = @ARGV;
$outfile = $title.'.'.$type;

#add or remove colours accordingly    
@colour_array = ("lblue", "brown", "blue", "lred", "lgreen", "yellow",
+ "green", "red", "purple", "orange", "pink", "dyellow");

@data;
$counter;

open IN, "$infile" or die $!; 
$label_line = <IN>; chomp $label_line;    #find the labels data (for n
+aming the colours in the stacked bars)
@split_labels = split(/\t/, $label_line);

while(<IN>)
{
    chomp; 
      $counter = 0;
      for (split/\t/,$_)
    { push @{$data[$counter++]}, $_; }
}
close IN;

#x-axis labels is the row #0 in the AoA @data
@x = @{$data[0]};

$total_y = @data-1;    #find how many y rows we have => how many colou
+rs we need (-1 because #0 is x-axis)
@pick_colours = @colour_array[ 0 .. $total_y - 1 ];

# Initiate the chart object 
$chart = Chart::Gnuplot->new( 
    output => $outfile, 
    title => $title, 
    xlabel => { 
            text => 'Method', 
            color => 'black',
            offset => "-1",
            font => 'Times-Roman, 20'
          },
    ylabel => { 
            text => $plot_title, 
            color => 'black',
            offset => -1,
            font => 'Times-Roman, 20'
          },
    yrange => [0, '*'], 
    xtics => { rotate => '-270', font => 'Times-Roman, 15' },    
    ytics => { font => 'Times-Roman, 15' },    
    bg => {
            color   => 'white',
            density => 0.2,
          },
    plotbg => 'white',
    legend => {position => 'outside bottom'}, 
    border => {
                sides    => 'bottom, left',
                linetype => 3,
                width    => 2,
                color    => 'black',
          },

    "style histogram" => 'rowstacked'
); 

#all the rest rows in @data are used for creating the histogram bars
#create object foreach series of data (column in excel)
$counter_objects=0;
for $y ( 1 .. $#data ) 
{
    $legend_name = $split_labels[$y];    #get the name for the legend,
+ e.g 1TM, 2TM, >2TM
    @tmp_y = @{$data[$y]};            #create temp array

    $counter_objects++;
    
    ${y."$counter_objects"} = Chart::Gnuplot::DataSet->new
    ( 
        xdata => \@x, 
        ydata => \@tmp_y, 
        title => "$legend_name", 
        fill => {density => 0.2}, 
        style => "histograms", 
    ); 
}
#$chart->plot2d($h1, $h2, $h3, $h4);


My main problem is that I want this script to be a general one, so I cannot know beforehand how many rows and columns I will have... I based my script in one example provided on the net, but in that one, the user had 4 arrays ( columns), which were hardcoded, so it was easy to say $chart -> plot2d ($h1, $h2, $3, $4). How can I do it in my case? And also, is the way I am trying to create the objects in the above for loop dynamically, correct? I think now, but I cannot figure out which is the correct syntax in this case. My thought was to somehow create the objects for the chart (in this case I need 4, aka 0TM, 1TM, 2TM, >2TM).

[poj]

It looks like you are trying to create a stacked bar histogram. Do you want horizontal or vertical bars ?

poj

[hermes1908]

Your code has quite a few problems and it seems that you are not grasping some fundamental concepts. If you are new to the language I would recommend you pick up a copy of "learning perl". I'm too lazy to comment each line of your code, but I'm sure some kind monk will. A thing which you have to be careful about this particular file is that it uses crlf (windows style) newlines. As I understand it perl automatically accounts for this on windows, but if you're on unix you need to explicitly open the file with "<:crlf".
Some important concepts you should familiarize yourself with:
• Variable references vs Normal variables (perldata)
• Scalar vs List context (perldata)
• Looping constructs in perl (perlsyn)
• Default variables (perlvar)

The list is by no means complete and as mentioned I would recommend getting a good book perl (learning perl or programming perl)

Below is some code which does what you want using what I believe are reasonably good conventions.


Good Luck
Hermes

#!/usr/bin/env perl

#Good practice (prevents autovivification)
use strict; 

my (@x, @y1, @y2, @y3, @y4);

open FILE, '<:crlf', "test.txt";

#Get and store header data
chomp(my $headerline=<FILE>);
my @headerdata=split /\t/, $headerline;


#Process each subsequent line of the file, storing the current line in
+ $_
while(<FILE>) { 
    #Remove newline from $_ (var holding current line)
    chomp; 

    #Split $_ using tab as delimeter and store contents in @rowdata
    my @rowdata=split /\t/; 

    #Store first column of the line in @x, etc..
    push @x, shift @rowdata; 
    push @y1, shift @rowdata; 
    push @y2, shift @rowdata; 
    push @y3, shift @rowdata; 
    push @y4, shift @rowdata; 
}
close FILE;

print "Array headerdata: @headerdata\n";
print "Array x: @x\n";
print "Array y1: @y1\n";
print "Array y2: @y2\n";
print "Array y3: @y3\n";
print "Array y4: @y4\n";

[poj]

You'll be close if you change this line

#$method_line = $_;    #lines with the methods/results for each one
$method_line = $lines[$i];
chomp($method_line);

and this

#unshift @AoA, [ @methods ];
unshift @AoA, [ @split_header ];

poj

[code-ninja]

The simplest way to do this can be:

use strict;
use warnings;

my @AoA;

while(<STREAM>){

chomp;
next if (/METHOD/);
push @AoA, [split /\s+/];

}

print \@AoA;

__DATA__
METHOD        0TM        1TM        2TM        >2TM
method1        3        68        18        10
method2        3        80        10        6
method3        3        87        9        1
method4        5        83        9        3
method5        6        75        15        4
method6        14        77        6        3
method7        12        82        2        3
method8        9        84        3        3
method9        13        68        16        3
method10    5        65        20        9
method11    9        64        20        7
method12    5        68        18        9
method13    6        82        9        2
method14    11        79        8        2
method15    6        69        20        5
method16    5        80        11        3
method17    1        68        17        14

What happens here is, the split "tokenizes" each line assuming the space between the words as the delimiter. The square brackets around the split create a reference to an anonymous array which is pushed into the array of arrays data structure.

UPDATE

Umm 2teez already said what I wrote above... I did not check the replies before posting... mesa bad.