3 perlref - Perl references and nested data structures
7 Before release 5 of Perl it was difficult to represent complex data
8 structures, because all references had to be symbolic, and even that was
9 difficult to do when you wanted to refer to a variable rather than a
10 symbol table entry. Perl not only makes it easier to use symbolic
11 references to variables, but lets you have "hard" references to any piece
12 of data. Any scalar may hold a hard reference. Because arrays and hashes
13 contain scalars, you can now easily build arrays of arrays, arrays of
14 hashes, hashes of arrays, arrays of hashes of functions, and so on.
16 Hard references are smart--they keep track of reference counts for you,
17 automatically freeing the thing referred to when its reference count
18 goes to zero. (Note: The reference counts for values in self-referential
19 or cyclic data structures may not go to zero without a little help; see
20 L<perlobj/"Two-Phased Garbage Collection"> for a detailed explanation.
21 If that thing happens to be an object, the object is
22 destructed. See L<perlobj> for more about objects. (In a sense,
23 everything in Perl is an object, but we usually reserve the word for
24 references to objects that have been officially "blessed" into a class package.)
26 A symbolic reference contains the name of a variable, just as a
27 symbolic link in the filesystem contains merely the name of a file.
28 The C<*glob> notation is a kind of symbolic reference. Hard references
29 are more like hard links in the file system: merely another way
30 at getting at the same underlying object, irrespective of its name.
32 "Hard" references are easy to use in Perl. There is just one
33 overriding principle: Perl does no implicit referencing or
34 dereferencing. When a scalar is holding a reference, it always behaves
35 as a scalar. It doesn't magically start being an array or a hash
36 unless you tell it so explicitly by dereferencing it.
38 References can be constructed in several ways.
44 By using the backslash operator on a variable, subroutine, or value.
45 (This works much like the & (address-of) operator in C.) Note
46 that this typically creates I<ANOTHER> reference to a variable, because
47 there's already a reference to the variable in the symbol table. But
48 the symbol table reference might go away, and you'll still have the
49 reference that the backslash returned. Here are some examples:
57 It isn't possible to create a true reference to an IO handle (filehandle or
58 dirhandle) using the backslash operator. See the explanation of the
59 *foo{THING} syntax below. (However, you're apt to find Perl code
60 out there using globrefs as though they were IO handles, which is
61 grandfathered into continued functioning.)
65 A reference to an anonymous array can be constructed using square
68 $arrayref = [1, 2, ['a', 'b', 'c']];
70 Here we've constructed a reference to an anonymous array of three elements
71 whose final element is itself a reference to another anonymous array of three
72 elements. (The multidimensional syntax described later can be used to
73 access this. For example, after the above, C<$arrayref-E<gt>[2][1]> would have
76 Note that taking a reference to an enumerated list is not the same
77 as using square brackets--instead it's the same as creating
80 @list = (\$a, \@b, \%c);
81 @list = \($a, @b, %c); # same thing!
83 As a special case, C<\(@foo)> returns a list of references to the contents
84 of C<@foo>, not a reference to C<@foo> itself. Likewise for C<%foo>.
88 A reference to an anonymous hash can be constructed using curly
96 Anonymous hash and array constructors can be intermixed freely to
97 produce as complicated a structure as you want. The multidimensional
98 syntax described below works for these too. The values above are
99 literals, but variables and expressions would work just as well, because
100 assignment operators in Perl (even within local() or my()) are executable
101 statements, not compile-time declarations.
103 Because curly brackets (braces) are used for several other things
104 including BLOCKs, you may occasionally have to disambiguate braces at the
105 beginning of a statement by putting a C<+> or a C<return> in front so
106 that Perl realizes the opening brace isn't starting a BLOCK. The economy and
107 mnemonic value of using curlies is deemed worth this occasional extra
110 For example, if you wanted a function to make a new hash and return a
111 reference to it, you have these options:
113 sub hashem { { @_ } } # silently wrong
114 sub hashem { +{ @_ } } # ok
115 sub hashem { return { @_ } } # ok
119 A reference to an anonymous subroutine can be constructed by using
120 C<sub> without a subname:
122 $coderef = sub { print "Boink!\n" };
124 Note the presence of the semicolon. Except for the fact that the code
125 inside isn't executed immediately, a C<sub {}> is not so much a
126 declaration as it is an operator, like C<do{}> or C<eval{}>. (However, no
127 matter how many times you execute that line (unless you're in an
128 C<eval("...")>), C<$coderef> will still have a reference to the I<SAME>
129 anonymous subroutine.)
131 Anonymous subroutines act as closures with respect to my() variables,
132 that is, variables visible lexically within the current scope. Closure
133 is a notion out of the Lisp world that says if you define an anonymous
134 function in a particular lexical context, it pretends to run in that
135 context even when it's called outside of the context.
137 In human terms, it's a funny way of passing arguments to a subroutine when
138 you define it as well as when you call it. It's useful for setting up
139 little bits of code to run later, such as callbacks. You can even
140 do object-oriented stuff with it, though Perl already provides a different
141 mechanism to do that--see L<perlobj>.
143 You can also think of closure as a way to write a subroutine template without
144 using eval. (In fact, in version 5.000, eval was the I<only> way to get
145 closures. You may wish to use "require 5.001" if you use closures.)
147 Here's a small example of how closures works:
151 return sub { my $y = shift; print "$x, $y!\n"; };
153 $h = newprint("Howdy");
154 $g = newprint("Greetings");
164 Greetings, earthlings!
166 Note particularly that $x continues to refer to the value passed into
167 newprint() I<despite> the fact that the "my $x" has seemingly gone out of
168 scope by the time the anonymous subroutine runs. That's what closure
171 This applies to only lexical variables, by the way. Dynamic variables
172 continue to work as they have always worked. Closure is not something
173 that most Perl programmers need trouble themselves about to begin with.
177 References are often returned by special subroutines called constructors.
178 Perl objects are just references to a special kind of object that happens to know
179 which package it's associated with. Constructors are just special
180 subroutines that know how to create that association. They do so by
181 starting with an ordinary reference, and it remains an ordinary reference
182 even while it's also being an object. Constructors are customarily
183 named new(), but don't have to be:
185 $objref = new Doggie (Tail => 'short', Ears => 'long');
189 References of the appropriate type can spring into existence if you
190 dereference them in a context that assumes they exist. Because we haven't
191 talked about dereferencing yet, we can't show you any examples yet.
195 A reference can be created by using a special syntax, lovingly known as
196 the *foo{THING} syntax. *foo{THING} returns a reference to the THING
197 slot in *foo (which is the symbol table entry which holds everything
200 $scalarref = *foo{SCALAR};
201 $arrayref = *ARGV{ARRAY};
202 $hashref = *ENV{HASH};
203 $coderef = *handler{CODE};
205 $globref = *foo{GLOB};
207 All of these are self-explanatory except for *foo{IO}. It returns the
208 IO handle, used for file handles (L<perlfunc/open>), sockets
209 (L<perlfunc/socket> and L<perlfunc/socketpair>), and directory handles
210 (L<perlfunc/opendir>). For compatibility with previous versions of
211 Perl, *foo{FILEHANDLE} is a synonym for *foo{IO}.
213 *foo{THING} returns undef if that particular THING hasn't been used yet,
214 except in the case of scalars. *foo{SCALAR} returns a reference to an
215 anonymous scalar if $foo hasn't been used yet. This might change in a
218 The use of *foo{IO} is the best way to pass bareword filehandles into or
219 out of subroutines, or to store them in larger data structures.
221 splutter(*STDOUT{IO});
224 print $fh "her um well a hmmm\n";
227 $rec = get_rec(*STDIN{IO});
233 Beware, though, that you can't do this with a routine which is going to
234 open the filehandle for you, because *HANDLE{IO} will be undef if HANDLE
235 hasn't been used yet. Use \*HANDLE for that sort of thing instead.
237 Using \*HANDLE (or *HANDLE) is another way to use and store non-bareword
238 filehandles (before perl version 5.002 it was the only way). The two
239 methods are largely interchangeable, you can do
242 $rec = get_rec(\*STDIN);
244 with the above subroutine definitions.
248 That's it for creating references. By now you're probably dying to
249 know how to use references to get back to your long-lost data. There
250 are several basic methods.
256 Anywhere you'd put an identifier (or chain of identifiers) as part
257 of a variable or subroutine name, you can replace the identifier with
258 a simple scalar variable containing a reference of the correct type:
261 push(@$arrayref, $filename);
262 $$arrayref[0] = "January";
263 $$hashref{"KEY"} = "VALUE";
265 print $globref "output\n";
267 It's important to understand that we are specifically I<NOT> dereferencing
268 C<$arrayref[0]> or C<$hashref{"KEY"}> there. The dereference of the
269 scalar variable happens I<BEFORE> it does any key lookups. Anything more
270 complicated than a simple scalar variable must use methods 2 or 3 below.
271 However, a "simple scalar" includes an identifier that itself uses method
272 1 recursively. Therefore, the following prints "howdy".
274 $refrefref = \\\"howdy";
279 Anywhere you'd put an identifier (or chain of identifiers) as part of a
280 variable or subroutine name, you can replace the identifier with a
281 BLOCK returning a reference of the correct type. In other words, the
282 previous examples could be written like this:
284 $bar = ${$scalarref};
285 push(@{$arrayref}, $filename);
286 ${$arrayref}[0] = "January";
287 ${$hashref}{"KEY"} = "VALUE";
289 $globref->print("output\n"); # iff IO::Handle is loaded
291 Admittedly, it's a little silly to use the curlies in this case, but
292 the BLOCK can contain any arbitrary expression, in particular,
293 subscripted expressions:
295 &{ $dispatch{$index} }(1,2,3); # call correct routine
297 Because of being able to omit the curlies for the simple case of C<$$x>,
298 people often make the mistake of viewing the dereferencing symbols as
299 proper operators, and wonder about their precedence. If they were,
300 though, you could use parentheses instead of braces. That's not the case.
301 Consider the difference below; case 0 is a short-hand version of case 1,
304 $$hashref{"KEY"} = "VALUE"; # CASE 0
305 ${$hashref}{"KEY"} = "VALUE"; # CASE 1
306 ${$hashref{"KEY"}} = "VALUE"; # CASE 2
307 ${$hashref->{"KEY"}} = "VALUE"; # CASE 3
309 Case 2 is also deceptive in that you're accessing a variable
310 called %hashref, not dereferencing through $hashref to the hash
311 it's presumably referencing. That would be case 3.
315 The case of individual array elements arises often enough that it gets
316 cumbersome to use method 2. As a form of syntactic sugar, the two
317 lines like that above can be written:
319 $arrayref->[0] = "January";
320 $hashref->{"KEY"} = "VALUE";
322 The left side of the array can be any expression returning a reference,
323 including a previous dereference. Note that C<$array[$x]> is I<NOT> the
324 same thing as C<$array-E<gt>[$x]> here:
326 $array[$x]->{"foo"}->[0] = "January";
328 This is one of the cases we mentioned earlier in which references could
329 spring into existence when in an lvalue context. Before this
330 statement, C<$array[$x]> may have been undefined. If so, it's
331 automatically defined with a hash reference so that we can look up
332 C<{"foo"}> in it. Likewise C<$array[$x]-E<gt>{"foo"}> will automatically get
333 defined with an array reference so that we can look up C<[0]> in it.
335 One more thing here. The arrow is optional I<BETWEEN> brackets
336 subscripts, so you can shrink the above down to
338 $array[$x]{"foo"}[0] = "January";
340 Which, in the degenerate case of using only ordinary arrays, gives you
341 multidimensional arrays just like C's:
343 $score[$x][$y][$z] += 42;
345 Well, okay, not entirely like C's arrays, actually. C doesn't know how
346 to grow its arrays on demand. Perl does.
350 If a reference happens to be a reference to an object, then there are
351 probably methods to access the things referred to, and you should probably
352 stick to those methods unless you're in the class package that defines the
353 object's methods. In other words, be nice, and don't violate the object's
354 encapsulation without a very good reason. Perl does not enforce
355 encapsulation. We are not totalitarians here. We do expect some basic
360 The ref() operator may be used to determine what type of thing the
361 reference is pointing to. See L<perlfunc>.
363 The bless() operator may be used to associate a reference with a package
364 functioning as an object class. See L<perlobj>.
366 A typeglob may be dereferenced the same way a reference can, because
367 the dereference syntax always indicates the kind of reference desired.
368 So C<${*foo}> and C<${\$foo}> both indicate the same scalar variable.
370 Here's a trick for interpolating a subroutine call into a string:
372 print "My sub returned @{[mysub(1,2,3)]} that time.\n";
374 The way it works is that when the C<@{...}> is seen in the double-quoted
375 string, it's evaluated as a block. The block creates a reference to an
376 anonymous array containing the results of the call to C<mysub(1,2,3)>. So
377 the whole block returns a reference to an array, which is then
378 dereferenced by C<@{...}> and stuck into the double-quoted string. This
379 chicanery is also useful for arbitrary expressions:
381 print "That yields @{[$n + 5]} widgets\n";
383 =head2 Symbolic references
385 We said that references spring into existence as necessary if they are
386 undefined, but we didn't say what happens if a value used as a
387 reference is already defined, but I<ISN'T> a hard reference. If you
388 use it as a reference in this case, it'll be treated as a symbolic
389 reference. That is, the value of the scalar is taken to be the I<NAME>
390 of a variable, rather than a direct link to a (possibly) anonymous
393 People frequently expect it to work like this. So it does.
396 $$name = 1; # Sets $foo
397 ${$name} = 2; # Sets $foo
398 ${$name x 2} = 3; # Sets $foofoo
399 $name->[0] = 4; # Sets $foo[0]
400 @$name = (); # Clears @foo
401 &$name(); # Calls &foo() (as in Perl 4)
403 ${"${pack}::$name"} = 5; # Sets $THAT::foo without eval
405 This is very powerful, and slightly dangerous, in that it's possible
406 to intend (with the utmost sincerity) to use a hard reference, and
407 accidentally use a symbolic reference instead. To protect against
412 and then only hard references will be allowed for the rest of the enclosing
413 block. An inner block may countermand that with
417 Only package variables are visible to symbolic references. Lexical
418 variables (declared with my()) aren't in a symbol table, and thus are
419 invisible to this mechanism. For example:
428 This will still print 10, not 20. Remember that local() affects package
429 variables, which are all "global" to the package.
431 =head2 Not-so-symbolic references
433 A new feature contributing to readability in perl version 5.001 is that the
434 brackets around a symbolic reference behave more like quotes, just as they
435 always have within a string. That is,
440 has always meant to print "pop on over", despite the fact that push is
441 a reserved word. This has been generalized to work the same outside
444 print ${push} . "over";
448 print ${ push } . "over";
450 will have the same effect. (This would have been a syntax error in
451 Perl 5.000, though Perl 4 allowed it in the spaceless form.) Note that this
452 construct is I<not> considered to be a symbolic reference when you're
456 ${ bareword }; # Okay, means $bareword.
457 ${ "bareword" }; # Error, symbolic reference.
459 Similarly, because of all the subscripting that is done using single
460 words, we've applied the same rule to any bareword that is used for
461 subscripting a hash. So now, instead of writing
463 $array{ "aaa" }{ "bbb" }{ "ccc" }
467 $array{ aaa }{ bbb }{ ccc }
469 and not worry about whether the subscripts are reserved words. In the
470 rare event that you do wish to do something like
474 you can force interpretation as a reserved word by adding anything that
475 makes it more than a bareword:
481 The B<-w> switch will warn you if it interprets a reserved word as a string.
482 But it will no longer warn you about using lowercase words, because the
483 string is effectively quoted.
487 You may not (usefully) use a reference as the key to a hash. It will be
488 converted into a string:
492 If you try to dereference the key, it won't do a hard dereference, and
493 you won't accomplish what you're attempting. You might want to do something
499 And then at least you can use the values(), which will be
500 real refs, instead of the keys(), which won't.
504 Besides the obvious documents, source code can be instructive.
505 Some rather pathological examples of the use of references can be found
506 in the F<t/op/ref.t> regression test in the Perl source directory.
508 See also L<perldsc> and L<perllol> for how to use references to create
509 complex data structures, and L<perlobj> for how to use them to create