2 X<reference> X<pointer> X<data structure> X<structure> X<struct>
4 perlref - Perl references and nested data structures
8 This is complete documentation about all aspects of references.
9 For a shorter, tutorial introduction to just the essential features,
14 Before release 5 of Perl it was difficult to represent complex data
15 structures, because all references had to be symbolic--and even then
16 it was difficult to refer to a variable instead of a symbol table entry.
17 Perl now not only makes it easier to use symbolic references to variables,
18 but also lets you have "hard" references to any piece of data or code.
19 Any scalar may hold a hard reference. Because arrays and hashes contain
20 scalars, you can now easily build arrays of arrays, arrays of hashes,
21 hashes of arrays, arrays of hashes of functions, and so on.
23 Hard references are smart--they keep track of reference counts for you,
24 automatically freeing the thing referred to when its reference count goes
25 to zero. (Reference counts for values in self-referential or
26 cyclic data structures may not go to zero without a little help; see
27 L<perlobj/"Two-Phased Garbage Collection"> for a detailed explanation.)
28 If that thing happens to be an object, the object is destructed. See
29 L<perlobj> for more about objects. (In a sense, everything in Perl is an
30 object, but we usually reserve the word for references to objects that
31 have been officially "blessed" into a class package.)
33 Symbolic references are names of variables or other objects, just as a
34 symbolic link in a Unix filesystem contains merely the name of a file.
35 The C<*glob> notation is something of a symbolic reference. (Symbolic
36 references are sometimes called "soft references", but please don't call
37 them that; references are confusing enough without useless synonyms.)
38 X<reference, symbolic> X<reference, soft>
39 X<symbolic reference> X<soft reference>
41 In contrast, hard references are more like hard links in a Unix file
42 system: They are used to access an underlying object without concern for
43 what its (other) name is. When the word "reference" is used without an
44 adjective, as in the following paragraph, it is usually talking about a
46 X<reference, hard> X<hard reference>
48 References are easy to use in Perl. There is just one overriding
49 principle: Perl does no implicit referencing or dereferencing. When a
50 scalar is holding a reference, it always behaves as a simple scalar. It
51 doesn't magically start being an array or hash or subroutine; you have to
52 tell it explicitly to do so, by dereferencing it.
54 =head2 Making References
55 X<reference, creation> X<referencing>
57 References can be created in several ways.
64 By using the backslash operator on a variable, subroutine, or value.
65 (This works much like the & (address-of) operator in C.)
66 This typically creates I<another> reference to a variable, because
67 there's already a reference to the variable in the symbol table. But
68 the symbol table reference might go away, and you'll still have the
69 reference that the backslash returned. Here are some examples:
77 It isn't possible to create a true reference to an IO handle (filehandle
78 or dirhandle) using the backslash operator. The most you can get is a
79 reference to a typeglob, which is actually a complete symbol table entry.
80 But see the explanation of the C<*foo{THING}> syntax below. However,
81 you can still use type globs and globrefs as though they were IO handles.
84 X<array, anonymous> X<[> X<[]> X<square bracket>
85 X<bracket, square> X<arrayref> X<array reference> X<reference, array>
87 A reference to an anonymous array can be created using square
90 $arrayref = [1, 2, ['a', 'b', 'c']];
92 Here we've created a reference to an anonymous array of three elements
93 whose final element is itself a reference to another anonymous array of three
94 elements. (The multidimensional syntax described later can be used to
95 access this. For example, after the above, C<< $arrayref->[2][1] >> would have
98 Taking a reference to an enumerated list is not the same
99 as using square brackets--instead it's the same as creating
100 a list of references!
102 @list = (\$a, \@b, \%c);
103 @list = \($a, @b, %c); # same thing!
105 As a special case, C<\(@foo)> returns a list of references to the contents
106 of C<@foo>, not a reference to C<@foo> itself. Likewise for C<%foo>,
107 except that the key references are to copies (since the keys are just
108 strings rather than full-fledged scalars).
111 X<hash, anonymous> X<{> X<{}> X<curly bracket>
112 X<bracket, curly> X<brace> X<hashref> X<hash reference> X<reference, hash>
114 A reference to an anonymous hash can be created using curly
122 Anonymous hash and array composers like these can be intermixed freely to
123 produce as complicated a structure as you want. The multidimensional
124 syntax described below works for these too. The values above are
125 literals, but variables and expressions would work just as well, because
126 assignment operators in Perl (even within local() or my()) are executable
127 statements, not compile-time declarations.
129 Because curly brackets (braces) are used for several other things
130 including BLOCKs, you may occasionally have to disambiguate braces at the
131 beginning of a statement by putting a C<+> or a C<return> in front so
132 that Perl realizes the opening brace isn't starting a BLOCK. The economy and
133 mnemonic value of using curlies is deemed worth this occasional extra
136 For example, if you wanted a function to make a new hash and return a
137 reference to it, you have these options:
139 sub hashem { { @_ } } # silently wrong
140 sub hashem { +{ @_ } } # ok
141 sub hashem { return { @_ } } # ok
143 On the other hand, if you want the other meaning, you can do this:
145 sub showem { { @_ } } # ambiguous (currently ok, but may change)
146 sub showem { {; @_ } } # ok
147 sub showem { { return @_ } } # ok
149 The leading C<+{> and C<{;> always serve to disambiguate
150 the expression to mean either the HASH reference, or the BLOCK.
153 X<subroutine, anonymous> X<subroutine, reference> X<reference, subroutine>
154 X<scope, lexical> X<closure> X<lexical> X<lexical scope>
156 A reference to an anonymous subroutine can be created by using
157 C<sub> without a subname:
159 $coderef = sub { print "Boink!\n" };
161 Note the semicolon. Except for the code
162 inside not being immediately executed, a C<sub {}> is not so much a
163 declaration as it is an operator, like C<do{}> or C<eval{}>. (However, no
164 matter how many times you execute that particular line (unless you're in an
165 C<eval("...")>), $coderef will still have a reference to the I<same>
166 anonymous subroutine.)
168 Anonymous subroutines act as closures with respect to my() variables,
169 that is, variables lexically visible within the current scope. Closure
170 is a notion out of the Lisp world that says if you define an anonymous
171 function in a particular lexical context, it pretends to run in that
172 context even when it's called outside the context.
174 In human terms, it's a funny way of passing arguments to a subroutine when
175 you define it as well as when you call it. It's useful for setting up
176 little bits of code to run later, such as callbacks. You can even
177 do object-oriented stuff with it, though Perl already provides a different
178 mechanism to do that--see L<perlobj>.
180 You might also think of closure as a way to write a subroutine
181 template without using eval(). Here's a small example of how
186 return sub { my $y = shift; print "$x, $y!\n"; };
188 $h = newprint("Howdy");
189 $g = newprint("Greetings");
199 Greetings, earthlings!
201 Note particularly that $x continues to refer to the value passed
202 into newprint() I<despite> "my $x" having gone out of scope by the
203 time the anonymous subroutine runs. That's what a closure is all
206 This applies only to lexical variables, by the way. Dynamic variables
207 continue to work as they have always worked. Closure is not something
208 that most Perl programmers need trouble themselves about to begin with.
211 X<constructor> X<new>
213 References are often returned by special subroutines called constructors. Perl
214 objects are just references to a special type of object that happens to know
215 which package it's associated with. Constructors are just special subroutines
216 that know how to create that association. They do so by starting with an
217 ordinary reference, and it remains an ordinary reference even while it's also
218 being an object. Constructors are often named C<new()>. You I<can> call them
221 $objref = new Doggie( Tail => 'short', Ears => 'long' );
223 But that can produce ambiguous syntax in certain cases, so it's often
224 better to use the direct method invocation approach:
226 $objref = Doggie->new(Tail => 'short', Ears => 'long');
229 $terminal = Term::Cap->Tgetent( { OSPEED => 9600 });
232 $main = MainWindow->new();
233 $menubar = $main->Frame(-relief => "raised",
239 References of the appropriate type can spring into existence if you
240 dereference them in a context that assumes they exist. Because we haven't
241 talked about dereferencing yet, we can't show you any examples yet.
246 A reference can be created by using a special syntax, lovingly known as
247 the *foo{THING} syntax. *foo{THING} returns a reference to the THING
248 slot in *foo (which is the symbol table entry which holds everything
251 $scalarref = *foo{SCALAR};
252 $arrayref = *ARGV{ARRAY};
253 $hashref = *ENV{HASH};
254 $coderef = *handler{CODE};
256 $globref = *foo{GLOB};
257 $formatref = *foo{FORMAT};
259 All of these are self-explanatory except for C<*foo{IO}>. It returns
260 the IO handle, used for file handles (L<perlfunc/open>), sockets
261 (L<perlfunc/socket> and L<perlfunc/socketpair>), and directory
262 handles (L<perlfunc/opendir>). For compatibility with previous
263 versions of Perl, C<*foo{FILEHANDLE}> is a synonym for C<*foo{IO}>, though it
264 is deprecated as of 5.8.0. If deprecation warnings are in effect, it will warn
267 C<*foo{THING}> returns undef if that particular THING hasn't been used yet,
268 except in the case of scalars. C<*foo{SCALAR}> returns a reference to an
269 anonymous scalar if $foo hasn't been used yet. This might change in a
272 C<*foo{IO}> is an alternative to the C<*HANDLE> mechanism given in
273 L<perldata/"Typeglobs and Filehandles"> for passing filehandles
274 into or out of subroutines, or storing into larger data structures.
275 Its disadvantage is that it won't create a new filehandle for you.
276 Its advantage is that you have less risk of clobbering more than
277 you want to with a typeglob assignment. (It still conflates file
278 and directory handles, though.) However, if you assign the incoming
279 value to a scalar instead of a typeglob as we do in the examples
280 below, there's no risk of that happening.
282 splutter(*STDOUT); # pass the whole glob
283 splutter(*STDOUT{IO}); # pass both file and dir handles
287 print $fh "her um well a hmmm\n";
290 $rec = get_rec(*STDIN); # pass the whole glob
291 $rec = get_rec(*STDIN{IO}); # pass both file and dir handles
300 =head2 Using References
301 X<reference, use> X<dereferencing> X<dereference>
303 That's it for creating references. By now you're probably dying to
304 know how to use references to get back to your long-lost data. There
305 are several basic methods.
311 Anywhere you'd put an identifier (or chain of identifiers) as part
312 of a variable or subroutine name, you can replace the identifier with
313 a simple scalar variable containing a reference of the correct type:
316 push(@$arrayref, $filename);
317 $$arrayref[0] = "January";
318 $$hashref{"KEY"} = "VALUE";
320 print $globref "output\n";
322 It's important to understand that we are specifically I<not> dereferencing
323 C<$arrayref[0]> or C<$hashref{"KEY"}> there. The dereference of the
324 scalar variable happens I<before> it does any key lookups. Anything more
325 complicated than a simple scalar variable must use methods 2 or 3 below.
326 However, a "simple scalar" includes an identifier that itself uses method
327 1 recursively. Therefore, the following prints "howdy".
329 $refrefref = \\\"howdy";
335 Anywhere you'd put an identifier (or chain of identifiers) as part of a
336 variable or subroutine name, you can replace the identifier with a
337 BLOCK returning a reference of the correct type. In other words, the
338 previous examples could be written like this:
340 $bar = ${$scalarref};
341 push(@{$arrayref}, $filename);
342 ${$arrayref}[0] = "January";
343 ${$hashref}{"KEY"} = "VALUE";
345 $globref->print("output\n"); # iff IO::Handle is loaded
347 Admittedly, it's a little silly to use the curlies in this case, but
348 the BLOCK can contain any arbitrary expression, in particular,
349 subscripted expressions:
351 &{ $dispatch{$index} }(1,2,3); # call correct routine
353 Because of being able to omit the curlies for the simple case of C<$$x>,
354 people often make the mistake of viewing the dereferencing symbols as
355 proper operators, and wonder about their precedence. If they were,
356 though, you could use parentheses instead of braces. That's not the case.
357 Consider the difference below; case 0 is a short-hand version of case 1,
360 $$hashref{"KEY"} = "VALUE"; # CASE 0
361 ${$hashref}{"KEY"} = "VALUE"; # CASE 1
362 ${$hashref{"KEY"}} = "VALUE"; # CASE 2
363 ${$hashref->{"KEY"}} = "VALUE"; # CASE 3
365 Case 2 is also deceptive in that you're accessing a variable
366 called %hashref, not dereferencing through $hashref to the hash
367 it's presumably referencing. That would be case 3.
370 X<autovivification> X<< -> >> X<arrow>
372 Subroutine calls and lookups of individual array elements arise often
373 enough that it gets cumbersome to use method 2. As a form of
374 syntactic sugar, the examples for method 2 may be written:
376 $arrayref->[0] = "January"; # Array element
377 $hashref->{"KEY"} = "VALUE"; # Hash element
378 $coderef->(1,2,3); # Subroutine call
380 The left side of the arrow can be any expression returning a reference,
381 including a previous dereference. Note that C<$array[$x]> is I<not> the
382 same thing as C<< $array->[$x] >> here:
384 $array[$x]->{"foo"}->[0] = "January";
386 This is one of the cases we mentioned earlier in which references could
387 spring into existence when in an lvalue context. Before this
388 statement, C<$array[$x]> may have been undefined. If so, it's
389 automatically defined with a hash reference so that we can look up
390 C<{"foo"}> in it. Likewise C<< $array[$x]->{"foo"} >> will automatically get
391 defined with an array reference so that we can look up C<[0]> in it.
392 This process is called I<autovivification>.
394 One more thing here. The arrow is optional I<between> brackets
395 subscripts, so you can shrink the above down to
397 $array[$x]{"foo"}[0] = "January";
399 Which, in the degenerate case of using only ordinary arrays, gives you
400 multidimensional arrays just like C's:
402 $score[$x][$y][$z] += 42;
404 Well, okay, not entirely like C's arrays, actually. C doesn't know how
405 to grow its arrays on demand. Perl does.
410 If a reference happens to be a reference to an object, then there are
411 probably methods to access the things referred to, and you should probably
412 stick to those methods unless you're in the class package that defines the
413 object's methods. In other words, be nice, and don't violate the object's
414 encapsulation without a very good reason. Perl does not enforce
415 encapsulation. We are not totalitarians here. We do expect some basic
420 Using a string or number as a reference produces a symbolic reference,
421 as explained above. Using a reference as a number produces an
422 integer representing its storage location in memory. The only
423 useful thing to be done with this is to compare two references
424 numerically to see whether they refer to the same location.
425 X<reference, numeric context>
427 if ($ref1 == $ref2) { # cheap numeric compare of references
428 print "refs 1 and 2 refer to the same thing\n";
431 Using a reference as a string produces both its referent's type,
432 including any package blessing as described in L<perlobj>, as well
433 as the numeric address expressed in hex. The ref() operator returns
434 just the type of thing the reference is pointing to, without the
435 address. See L<perlfunc/ref> for details and examples of its use.
436 X<reference, string context>
438 The bless() operator may be used to associate the object a reference
439 points to with a package functioning as an object class. See L<perlobj>.
441 A typeglob may be dereferenced the same way a reference can, because
442 the dereference syntax always indicates the type of reference desired.
443 So C<${*foo}> and C<${\$foo}> both indicate the same scalar variable.
445 Here's a trick for interpolating a subroutine call into a string:
447 print "My sub returned @{[mysub(1,2,3)]} that time.\n";
449 The way it works is that when the C<@{...}> is seen in the double-quoted
450 string, it's evaluated as a block. The block creates a reference to an
451 anonymous array containing the results of the call to C<mysub(1,2,3)>. So
452 the whole block returns a reference to an array, which is then
453 dereferenced by C<@{...}> and stuck into the double-quoted string. This
454 chicanery is also useful for arbitrary expressions:
456 print "That yields @{[$n + 5]} widgets\n";
458 =head2 Symbolic references
459 X<reference, symbolic> X<reference, soft>
460 X<symbolic reference> X<soft reference>
462 We said that references spring into existence as necessary if they are
463 undefined, but we didn't say what happens if a value used as a
464 reference is already defined, but I<isn't> a hard reference. If you
465 use it as a reference, it'll be treated as a symbolic
466 reference. That is, the value of the scalar is taken to be the I<name>
467 of a variable, rather than a direct link to a (possibly) anonymous
470 People frequently expect it to work like this. So it does.
473 $$name = 1; # Sets $foo
474 ${$name} = 2; # Sets $foo
475 ${$name x 2} = 3; # Sets $foofoo
476 $name->[0] = 4; # Sets $foo[0]
477 @$name = (); # Clears @foo
478 &$name(); # Calls &foo() (as in Perl 4)
480 ${"${pack}::$name"} = 5; # Sets $THAT::foo without eval
482 This is powerful, and slightly dangerous, in that it's possible
483 to intend (with the utmost sincerity) to use a hard reference, and
484 accidentally use a symbolic reference instead. To protect against
489 and then only hard references will be allowed for the rest of the enclosing
490 block. An inner block may countermand that with
494 Only package variables (globals, even if localized) are visible to
495 symbolic references. Lexical variables (declared with my()) aren't in
496 a symbol table, and thus are invisible to this mechanism. For example:
505 This will still print 10, not 20. Remember that local() affects package
506 variables, which are all "global" to the package.
508 =head2 Not-so-symbolic references
510 A new feature contributing to readability in perl version 5.001 is that the
511 brackets around a symbolic reference behave more like quotes, just as they
512 always have within a string. That is,
517 has always meant to print "pop on over", even though push is
518 a reserved word. This has been generalized to work the same outside
521 print ${push} . "over";
525 print ${ push } . "over";
527 will have the same effect. (This would have been a syntax error in
528 Perl 5.000, though Perl 4 allowed it in the spaceless form.) This
529 construct is I<not> considered to be a symbolic reference when you're
533 ${ bareword }; # Okay, means $bareword.
534 ${ "bareword" }; # Error, symbolic reference.
536 Similarly, because of all the subscripting that is done using single
537 words, we've applied the same rule to any bareword that is used for
538 subscripting a hash. So now, instead of writing
540 $array{ "aaa" }{ "bbb" }{ "ccc" }
544 $array{ aaa }{ bbb }{ ccc }
546 and not worry about whether the subscripts are reserved words. In the
547 rare event that you do wish to do something like
551 you can force interpretation as a reserved word by adding anything that
552 makes it more than a bareword:
558 The C<use warnings> pragma or the B<-w> switch will warn you if it
559 interprets a reserved word as a string.
560 But it will no longer warn you about using lowercase words, because the
561 string is effectively quoted.
563 =head2 Pseudo-hashes: Using an array as a hash
564 X<pseudo-hash> X<pseudo hash> X<pseudohash>
566 Pseudo-hashes have been removed from Perl. The 'fields' pragma
569 =head2 Function Templates
570 X<scope, lexical> X<closure> X<lexical> X<lexical scope>
571 X<subroutine, nested> X<sub, nested> X<subroutine, local> X<sub, local>
573 As explained above, an anonymous function with access to the lexical
574 variables visible when that function was compiled, creates a closure. It
575 retains access to those variables even though it doesn't get run until
576 later, such as in a signal handler or a Tk callback.
578 Using a closure as a function template allows us to generate many functions
579 that act similarly. Suppose you wanted functions named after the colors
580 that generated HTML font changes for the various colors:
582 print "Be ", red("careful"), "with that ", green("light");
584 The red() and green() functions would be similar. To create these,
585 we'll assign a closure to a typeglob of the name of the function we're
588 @colors = qw(red blue green yellow orange purple violet);
589 for my $name (@colors) {
590 no strict 'refs'; # allow symbol table manipulation
591 *$name = *{uc $name} = sub { "<FONT COLOR='$name'>@_</FONT>" };
594 Now all those different functions appear to exist independently. You can
595 call red(), RED(), blue(), BLUE(), green(), etc. This technique saves on
596 both compile time and memory use, and is less error-prone as well, since
597 syntax checks happen at compile time. It's critical that any variables in
598 the anonymous subroutine be lexicals in order to create a proper closure.
599 That's the reasons for the C<my> on the loop iteration variable.
601 This is one of the only places where giving a prototype to a closure makes
602 much sense. If you wanted to impose scalar context on the arguments of
603 these functions (probably not a wise idea for this particular example),
604 you could have written it this way instead:
606 *$name = sub ($) { "<FONT COLOR='$name'>$_[0]</FONT>" };
608 However, since prototype checking happens at compile time, the assignment
609 above happens too late to be of much use. You could address this by
610 putting the whole loop of assignments within a BEGIN block, forcing it
611 to occur during compilation.
613 Access to lexicals that change over type--like those in the C<for> loop
614 above--only works with closures, not general subroutines. In the general
615 case, then, named subroutines do not nest properly, although anonymous
616 ones do. Thus is because named subroutines are created (and capture any
617 outer lexicals) only once at compile time, whereas anonymous subroutines
618 get to capture each time you execute the 'sub' operator. If you are
619 accustomed to using nested subroutines in other programming languages with
620 their own private variables, you'll have to work at it a bit in Perl. The
621 intuitive coding of this type of thing incurs mysterious warnings about
622 "will not stay shared". For example, this won't work:
626 sub inner { return $x * 19 } # WRONG
630 A work-around is the following:
634 local *inner = sub { return $x * 19 };
638 Now inner() can only be called from within outer(), because of the
639 temporary assignments of the closure (anonymous subroutine). But when
640 it does, it has normal access to the lexical variable $x from the scope
643 This has the interesting effect of creating a function local to another
644 function, something not normally supported in Perl.
647 X<reference, string context> X<reference, use as hash key>
649 You may not (usefully) use a reference as the key to a hash. It will be
650 converted into a string:
654 If you try to dereference the key, it won't do a hard dereference, and
655 you won't accomplish what you're attempting. You might want to do something
661 And then at least you can use the values(), which will be
662 real refs, instead of the keys(), which won't.
664 The standard Tie::RefHash module provides a convenient workaround to this.
668 Besides the obvious documents, source code can be instructive.
669 Some pathological examples of the use of references can be found
670 in the F<t/op/ref.t> regression test in the Perl source directory.
672 See also L<perldsc> and L<perllol> for how to use references to create
673 complex data structures, and L<perltoot>, L<perlobj>, and L<perlbot>
674 for how to use them to create objects.