perl 中的引用：数组到另一个数组的散列

Question

I have a problem with referencing a hash in an array to another array. I have an array @result which looks like this:

@result = (

{ "type" => "variable",
  "s" => "NGDP",
  "variable" => "NGDP" },

{"type" => "subject",
  "s" => "USA",
  "subject" => "USA",
  "variable" => "NGDP" },

{ "type" => "colon",
  "s" => ",",
  "colon" => "," }, 

{ "type" => "subject",
  "s" => "JPN",
  "subject" => "JPN",
  "variable" => "NGDP" },

{ "type" => "operator",
  "s" => "+",
  "operator => "+" },

{"type" => "subject",
  "s" => "CHN",
  "subject" => "CHN",
  "variable" => "NGDP" },
);

I want to divide this array into colons and push elements of the @result array to another array, so i wrote the script:

for ($i = 0; $i <= $#result; $i++) {  

if (defined $result[$i]{subject} or $result[$i]{operator} and not defined $result[$i]{colon}) {
  push @part_col, \%{$result[$i]};
}

elsif ($i == $#result) {
  push @part_col_all, \@part_col;
}

elsif (defined $result[$i]{colon}) {
  push @part_col_all, \@part_col;
  my @part_col;
 }
}

So what I need is that if I print out $part_col_all[0][0]{subject} the result will be "USA" ,

and for $part_col_all[1][0]{subject} will be "JPN" ,

and for $part_col_all[1][1]{operator} will be "+" etc.

My result for $part_col_all[0][0]{subject} is "USA"

and for $part_col_all[0][1]{subject} is "JPN" which should be in $part_col_all[1][0]{subject} .

The result for $part_col_all[0][3]{subject} is "CHN" , while it should be in $part_col_all[1][2]{subject}.

I'm making an application which is creating graphs from economical data based on a certain economical input. The @result array is my preprocessed input where I know to which country which variable belongs. If I get an input like GDP USA CAN, JPN+CHN I need to split this input to GDP USA CAN and JPN+CHN. That's why I made a condition, if colon is found, push everything in @part_col to the first element of @part_col_all, and then if it's on the end of the input, push JPN+CHN to the second element of @push_col_all.

So @part_col_all should looks like this:

@part_col_all = (
    (   
        {"type" => "subject",
        "s" => "USA",
        "subject" => "USA",
        "variable" => "NGDP" },

        {"type" => "subject",
        "s" => "CAN",
        "subject" => "CAN",
        "variable" => "NGDP" },
    ),

    (
        { "type" => "subject",
        "s" => "JPN",
        "subject" => "JPN",
        "variable" => "NGDP" },

        { "type" => "operator",
        "s" => "+",
        "operator" => "+" },

        {"type" => "subject",
        "s" => "CHN",
        "subject" => "CHN",
        "variable" => "NGDP" },

    )
);

I dont know what I'm doing wrong. Sorry if there are any basic mistakes, im a beginner. Thanks a lot.

Answer 1

First, you're missing a quote:

{ "type" => "operator",
  "s" => "+",
  "operator" => "+" },
           ^ missing

As for printing, you can do the following:

foreach my $part (@part_col){
    print $part->{operator}."\n";
}

Or do whatever you want in the print cycle with the values

Answer 2

You should read the Perl Reference Tutorial to help you.

There's no sin in dereferencing to simplify your code:

my @part_col;
my @part_col_all;

for $i ( 0..$#array ) {
    my %hash = ${ $result[$i] };  # Make it easy on yourself. Dereference
    if ( defined $hash{subject} or defined $hash{operator} and not defined $hash{colon} ) {
        push @part_col, \%hash;  # or push, @par_col, $result[$i]
    }
}

Notice I changed the for from the three part setup you had to a cleaner and easier to understand way of stating it.

Looking closer at your data structure, I notice that $hash{type} will tell you whether or not $hash{operator} , $hash{subject} , or $hash{colon} is defined. Let's just use $hash{type} and simplify that if :

my @part_col;
my @part_col_all;

for my $i ( 0..$#array ) { 
    my %hash = ${ $result[$i] };  # Make it easy on yourself. Dereference
    if ( $hash{type} eq "subject" or $hash{type} eq "operator" ) {
        push @part_col, \%hash;  # or push, @par_col, $result[$i]
    }
}

In fact, since @array is just an array, I'll treat it like one. I'll use a simple for structure to go through each element of my array. Each element is a hash_reference, so:

for my $hash_ref ( @array ) {
    my %hash = %{ %hash_ref };
    if ( $hash{type} eq "subject" or $hash{type} eq "operator" ) {
        push @part_col, \%hash;
    }
}

And further simplification, I can dereference and talk about a particular element of my hash all at once by using the -> syntax:

for my $hash_ref ( @array ) {
    if ( $hash_ref->{type} eq "subject" or $hash_ref->{type} eq "operator" ) {
        push @part_col, $hash_ref;
    }
}

I'm trying to understand the rest of your code:

    elsif ($i == $#result) {
        push @part_col_all, \@part_col;
    }

    elsif (defined $hash_ref->{colon}) {
        push @part_col_all, \@part_col;
        my @part_col;
    }
}

These pushes of @part_col onto @part_col_all confuse me. Exactly what are you trying to store in @part_col_all ? Remember that \\@part_col is the location in memory where you're storing @part_col . You're pushing that same memory location over and over onto that hash, so you're storing the same reference over and over again. Is that really what you want? I doubt it.

You need to do is to decide exactly what your data structure really represents. A data structure should have a solid definition. What does the data structure @part_col_all represent? What does the data structure $part_col_all[$i] represent? What does the data structure $part_col_all[$i]->[$j] represent? Without knowing this, it's very hard to answer the rest of your question.

Are you storing elements where the type is colon in one array and everything else in another array? Or are you storing everything in one array, and in another array, storing everything that's not a type colon ?

Once I understand this, I can answer the rest of your question.

Addendum

Thank you for your reply, I will try that way and write my results. It is realy helpful. I updated my question with more information about data structure of @part_col_all. I hope that you understand what I'm trying to explain, if not I'll try it again.

---

If I understand what you're doing, someone enters in NGDP USA , JPN+CNA and that means you're comparing the NGDP between the United States vs. Japan and China combined.

It seems to me that you would want three separate variables:

• $parameter - What you are measuring. (GDP, etc.)
• @countries_set_1 - The first set of countries
• @countries_set_2 - The second set of countries which you're comparing against the first set.

And, what you call the colon (which we would call a comma in the US) as a separator between the first set of countries vs. the second set. Then, you'd simply go through a loop. It could be that the two arrays are merely two elements of the same array, and the sets of countries are array references. I imagine something like this:

@input = qw(GDP USA, JPN CHN); # Compare the GDP of the USA with Japan and China together
my $parameter = shift @input;  # Remove what you're measuring
my @country_sets;              # An array of arrays
my $set = 0                    # Which set you're on
for my $value ( @input ) {
    if ( $value eq "," ) {
       $set += 1;              # Next Set
       next;
    }
    push @{ $country_sets[$set] }, $input;
}

This would create a data structure like this:

@country_sets  = (
                     (
                          USA,
                     ),
                     (
                          JPN,
                          CHN,
                     ),
                 )

No need for the complex @results since you're only going to have a single operation (GDP, etc.) for all involved.

However, I think I see what you want. We'll go with an array of arrays. Here's what I had before:

for my $hash_ref ( @array ) {
    if ( $hash_ref->{type} eq "subject" or $hash_ref->{type} eq "operator" ) {
        push @part_col, $hash_ref;
    }
}

We'll combine that and the code I offered right above which splits the countries into two sets:

my @country_sets;              # An array of arrays
my $set = 0                    # Which set you're on
for my $country_ref ( @array ) {
    next if $country_ref->{type} eq "variable";  # We don't want variables
    if ( $country_ref{type} eq "colon" ) {       # Switch to the other country set
        set += 1;
        next;
    }
    push @{ $country_sets[$set] }, $country_ref;
}

The first few entries will go into $country_sets[0] which will be an array reference. After the colon (which won't be input into the set), the second set of countries will go into $country_sets[1] which will be an other array_ref to a reference of hashes:

• @country_sets - Contains the input information into two sets
• @country_sets[$x] - A particular set of countries (and possibly operator)
• @country_sets[$x]->[$y] - A Particular country or operator
• @country_sets[$x]->[$y]->{$key} - A particular value from a particular country

Where $x goes from 0 to 1 . This will give you something like this:

$country_sets[0] = (
                       {
                          "type" => "subject",
                          "s" => "USA",
                           "subject" => "USA",
                           "variable" => "NGDP",
                       },
                   )

$country_sets[1] = (
                      {
                         "type" => "subject",
                         "s" => "JPN",
                         "subject" => "JPN",
                         "variable" => "NGDP",
                      },

                      { 
                          "type" => "operator",
                          "s" => "+",
                          "operator => "+",
                      },
                      {
                          "type" => "subject",
                          "s" => "CHN",
                          "subject" => "CHN",
                          "variable" => "NGDP",
                      },
                 );