How to use an array reference to pairs of elements

2019-01-26 02:05发布

问题:

I am considering this answer which uses a single array reference of points, where a point is a reference to a two-element array. My original code of the question (function extract-crossing) uses two separate arrays $x and $y here which I call like this:

my @x = @{ $_[0] }; my @y = @{ $_[1] };
...
return extract_crossing(\@x, \@y);

The new code below based on the answer takes (x, y) and returns a single datatype, here x intercept points:

use strict; use warnings;    
use Math::Geometry::Planar qw(SegmentLineIntersection);
use Test::Exception;

sub x_intercepts {
   my ($points) = @_;

   die 'Must pass at least 2 points' unless @$points >= 2;

   my @intercepts;
   my @x_axis = ( [0, 0], [1, 0] );

   foreach my $i (0 .. $#$points - 1) {
     my $intersect = SegmentLineIntersection([@$points[$i,$i+1],@x_axis]);
     push @intercepts, $intersect if $intersect;
   }

   return \@intercepts;
}

which I try to call like this:

my @x = @{ $_[0] }; my @y = @{ $_[1] };  
...
my $masi = x_intercepts(\@x);
return $masi; 

However, the code does not make sense. I am confused about passing the "double array" to the x_intercepts() function.

How can you make the example code clearer to the original setting?

回答1:

If I am understanding the problem here, @ThisSuitIsBlackNot++ has written a function (x_intercepts which is available in the thread: Function to extract intersections of crossing lines on axes) that expects its argument to be a reference to a list of array references. The x_intercepts subroutine in turn uses a function from Math::Geometry::Planar which expects the points of a line segment to be passed as series of array references/anonymous arrays that contain the x,y values for each point.

Again - it is not entirely clear - but it seems your data is in two different arrays: one containing all the x values and one with the corresponding y values. Is this the case? If this is not correct please leave a comment and I will remove this answer.

If that is the source of your problem then you can "munge" or transform your data before you pass it to x_intercepts or - as @ThisSuitIsBlackNot suggests - you can rewrite the function. Here is an example of munging your existing data into an "@input_list" to pass to x_intercepts.

my @xs = qw/-1 1 3/;
my @ys = qw/-1 1 -1 /;

my @input_list ;
foreach my $i ( 0..$#ys ) {
   push @input_list,  [ $xs[$i], $ys[$i] ]  ;
}

my $intercept_list = x_intercepts(\@input_list)  ;
say join ",", @$_ for @$intercept_list ;

Adding the lines above to your script produces:

Output:

0,0
2,0

You have to be very careful doing this kind of thing and using tests to make sure you are passing the correctly transformed data in an expected way is a good idea.


I think a more general difficulty is that until you are familiar with perl it is sometimes tricky to easily see what sorts of values a subroutine is expecting, where they end up after they are passed in, and how to access them.

A solid grasp of perl data structures can help with that - for example I think what you are calling a "double array" or "double element" here is an "array of arrays". There are ways to make it easier to see where default arguments passed to a subroutine (in @_) are going (notice how @ThisSuitIsBlackNot has passed them to a nicely named array reference: "($points)"). Copious re-reading of perldocperbsub can help things seem more obvious.

References are key to understanding perl subroutines since to pass an array or hash to a subrouting you need to do so by references. If the argument passed x_intercepts is a list of two lists of anonymous arrays then when it is assigned to ($points), @$points->[0] @$points->[1] will be the arrays contain those lists.

I hope this helps and is not too basic (or incorrect). If @ThisSuitIsBlackNot finds the time to provide an answer you should accept it: some very useful examples have been provided.