1PERLDATA(1) Perl Programmers Reference Guide PERLDATA(1)
2
6 perldata - Perl data types
7
9 Variable names
10 Perl has three built-in data types: scalars, arrays of scalars, and
11 associative arrays of scalars, known as "hashes". A scalar is a single
12 string (of any size, limited only by the available memory), number, or
13 a reference to something (which will be discussed in perlref). Normal
14 arrays are ordered lists of scalars indexed by number, starting with 0.
15 Hashes are unordered collections of scalar values indexed by their
16 associated string key.
17 Values are usually referred to by name, or through a named reference.
19 The first character of the name tells you to what sort of data
20 structure it refers. The rest of the name tells you the particular
21 value to which it refers. Usually this name is a single identifier,
22 that is, a string beginning with a letter or underscore, and containing
23 letters, underscores, and digits. In some cases, it may be a chain of
24 identifiers, separated by "::" (or by the slightly archaic "'"); all
25 but the last are interpreted as names of packages, to locate the
26 namespace in which to look up the final identifier (see "Packages" in
27 perlmod for details). For a more in-depth discussion on identifiers,
28 see "Identifier parsing". It's possible to substitute for a simple
29 identifier, an expression that produces a reference to the value at
30 runtime. This is described in more detail below and in perlref.
31 Perl also has its own built-in variables whose names don't follow these
33 rules. They have strange names so they don't accidentally collide with
34 one of your normal variables. Strings that match parenthesized parts
35 of a regular expression are saved under names containing only digits
36 after the "$" (see perlop and perlre). In addition, several special
37 variables that provide windows into the inner working of Perl have
38 names containing punctuation characters. These are documented in
39 perlvar.
40 Scalar values are always named with '$', even when referring to a
42 scalar that is part of an array or a hash. The '$' symbol works
43 semantically like the English word "the" in that it indicates a single
44 value is expected.
45 $days # the simple scalar value "days"
47 $days[28] # the 29th element of array @days
48 $days{'Feb'} # the 'Feb' value from hash %days
49 $#days # the last index of array @days
50 Entire arrays (and slices of arrays and hashes) are denoted by '@',
52 which works much as the word "these" or "those" does in English, in
53 that it indicates multiple values are expected.
54 @days # ($days[0], $days[1],... $days[n])
56 @days[3,4,5] # same as ($days[3],$days[4],$days[5])
57 @days{'a','c'} # same as ($days{'a'},$days{'c'})
58 Entire hashes are denoted by '%':
60 %days # (key1, val1, key2, val2 ...)
62 In addition, subroutines are named with an initial '&', though this is
64 optional when unambiguous, just as the word "do" is often redundant in
65 English. Symbol table entries can be named with an initial '*', but
66 you don't really care about that yet (if ever :-).
67 Every variable type has its own namespace, as do several non-variable
69 identifiers. This means that you can, without fear of conflict, use
70 the same name for a scalar variable, an array, or a hash--or, for that
71 matter, for a filehandle, a directory handle, a subroutine name, a
72 format name, or a label. This means that $foo and @foo are two
73 different variables. It also means that $foo[1] is a part of @foo, not
74 a part of $foo. This may seem a bit weird, but that's okay, because it
75 is weird.
76 Because variable references always start with '$', '@', or '%', the
78 "reserved" words aren't in fact reserved with respect to variable
79 names. They are reserved with respect to labels and filehandles,
80 however, which don't have an initial special character. You can't have
81 a filehandle named "log", for instance. Hint: you could say
82 "open(LOG,'logfile')" rather than "open(log,'logfile')". Using
83 uppercase filehandles also improves readability and protects you from
84 conflict with future reserved words. Case is significant--"FOO",
85 "Foo", and "foo" are all different names. Names that start with a
86 letter or underscore may also contain digits and underscores.
87 It is possible to replace such an alphanumeric name with an expression
89 that returns a reference to the appropriate type. For a description of
90 this, see perlref.
91 Names that start with a digit may contain only more digits. Names that
93 do not start with a letter, underscore, digit or a caret are limited to
94 one character, e.g., $% or $$. (Most of these one character names
95 have a predefined significance to Perl. For instance, $$ is the
96 current process id. And all such names are reserved for Perl's
97 possible use.)
98 Identifier parsing
100 Up until Perl 5.18, the actual rules of what a valid identifier was
101 were a bit fuzzy. However, in general, anything defined here should
102 work on previous versions of Perl, while the opposite -- edge cases
103 that work in previous versions, but aren't defined here -- probably
104 won't work on newer versions. As an important side note, please note
105 that the following only applies to bareword identifiers as found in
106 Perl source code, not identifiers introduced through symbolic
107 references, which have much fewer restrictions. If working under the
108 effect of the "use utf8;" pragma, the following rules apply:
109 / (?[ ( \p{Word} & \p{XID_Start} ) + [_] ])
111 (?[ ( \p{Word} & \p{XID_Continue} ) ]) * /x
112 That is, a "start" character followed by any number of "continue"
114 characters. Perl requires every character in an identifier to also
115 match "\w" (this prevents some problematic cases); and Perl
116 additionally accepts identifier names beginning with an underscore.
117 If not under "use utf8", the source is treated as ASCII + 128 extra
119 generic characters, and identifiers should match
120 / (?aa) (?!\d) \w+ /x
122 That is, any word character in the ASCII range, as long as the first
124 character is not a digit.
125 There are two package separators in Perl: A double colon ("::") and a
127 single quote ("'"). Normal identifiers can start or end with a double
128 colon, and can contain several parts delimited by double colons.
129 Single quotes have similar rules, but with the exception that they are
130 not legal at the end of an identifier: That is, "$'foo" and "$foo'bar"
131 are legal, but "$foo'bar'" is not.
132 Additionally, if the identifier is preceded by a sigil -- that is, if
134 the identifier is part of a variable name -- it may optionally be
135 enclosed in braces.
136 While you can mix double colons with singles quotes, the quotes must
138 come after the colons: "$::::'foo" and "$foo::'bar" are legal, but
139 "$::'::foo" and "$foo'::bar" are not.
140 Put together, a grammar to match a basic identifier becomes
142 /
144 (?(DEFINE)
145 (?<variable>
146 (?&sigil)
147 (?:
148 (?&normal_identifier)
149 | \{ \s* (?&normal_identifier) \s* \}
150 )
151 )
152 (?<normal_identifier>
153 (?: :: )* '?
154 (?&basic_identifier)
155 (?: (?= (?: :: )+ '? | (?: :: )* ' ) (?&normal_identifier) )?
156 (?: :: )*
157 )
158 (?<basic_identifier>
159 # is use utf8 on?
160 (?(?{ (caller(0))[8] & $utf8::hint_bits })
161 (?&Perl_XIDS) (?&Perl_XIDC)*
162 | (?aa) (?!\d) \w+
163 )
164 )
165 (?<sigil> [&*\$\@\%])
166 (?<Perl_XIDS> (?[ ( \p{Word} & \p{XID_Start} ) + [_] ]) )
167 (?<Perl_XIDC> (?[ \p{Word} & \p{XID_Continue} ]) )
168 )
169 /x
170 Meanwhile, special identifiers don't follow the above rules; For the
172 most part, all of the identifiers in this category have a special
173 meaning given by Perl. Because they have special parsing rules, these
174 generally can't be fully-qualified. They come in six forms (but don't
175 use forms 5 and 6):
176 1. A sigil, followed solely by digits matching "\p{POSIX_Digit}", like
178 $0, $1, or $10000.
179 2. A sigil followed by a single character matching the
181 "\p{POSIX_Punct}" property, like $! or "%+", except the character
182 "{" doesn't work.
183 3. A sigil, followed by a caret and any one of the characters
185 "[][A-Z^_?\]", like $^V or $^].
186 4. Similar to the above, a sigil, followed by bareword text in braces,
188 where the first character is a caret. The next character is any
189 one of the characters "[][A-Z^_?\]", followed by ASCII word
190 characters. An example is "${^GLOBAL_PHASE}".
191 5. A sigil, followed by any single character in the range
193 "[\xA1-\xAC\xAE-\xFF]" when not under "use utf8". (Under
194 "use utf8", the normal identifier rules given earlier in this
195 section apply.) Use of non-graphic characters (the C1 controls,
196 the NO-BREAK SPACE, and the SOFT HYPHEN) has been disallowed since
197 v5.26.0. The use of the other characters is unwise, as these are
198 all reserved to have special meaning to Perl, and none of them
199 currently do have special meaning, though this could change without
200 notice.
201 Note that an implication of this form is that there are identifiers
203 only legal under "use utf8", and vice-versa, for example the
204 identifier "$etat" is legal under "use utf8", but is otherwise
205 considered to be the single character variable "$e" followed by the
206 bareword "tat", the combination of which is a syntax error.
207 6. This is a combination of the previous two forms. It is valid only
209 when not under "use utf8" (normal identifier rules apply when under
210 "use utf8"). The form is a sigil, followed by text in braces,
211 where the first character is any one of the characters in the range
212 "[\x80-\xFF]" followed by ASCII word characters up to the trailing
213 brace.
214 The same caveats as the previous form apply: The non-graphic
216 characters are no longer allowed with "use utf8", it is unwise to
217 use this form at all, and utf8ness makes a big difference.
218 Prior to Perl v5.24, non-graphical ASCII control characters were also
220 allowed in some situations; this had been deprecated since v5.20.
221 Context
223 The interpretation of operations and values in Perl sometimes depends
224 on the requirements of the context around the operation or value.
225 There are two major contexts: list and scalar. Certain operations
226 return list values in contexts wanting a list, and scalar values
227 otherwise. If this is true of an operation it will be mentioned in the
228 documentation for that operation. In other words, Perl overloads
229 certain operations based on whether the expected return value is
230 singular or plural. Some words in English work this way, like "fish"
231 and "sheep".
232 In a reciprocal fashion, an operation provides either a scalar or a
234 list context to each of its arguments. For example, if you say
235 int( <STDIN> )
237 the integer operation provides scalar context for the <> operator,
239 which responds by reading one line from STDIN and passing it back to
240 the integer operation, which will then find the integer value of that
241 line and return that. If, on the other hand, you say
242 sort( <STDIN> )
244 then the sort operation provides list context for <>, which will
246 proceed to read every line available up to the end of file, and pass
247 that list of lines back to the sort routine, which will then sort those
248 lines and return them as a list to whatever the context of the sort
249 was.
250 Assignment is a little bit special in that it uses its left argument to
252 determine the context for the right argument. Assignment to a scalar
253 evaluates the right-hand side in scalar context, while assignment to an
254 array or hash evaluates the righthand side in list context. Assignment
255 to a list (or slice, which is just a list anyway) also evaluates the
256 right-hand side in list context.
257 When you use the "use warnings" pragma or Perl's -w command-line
259 option, you may see warnings about useless uses of constants or
260 functions in "void context". Void context just means the value has
261 been discarded, such as a statement containing only ""fred";" or
262 "getpwuid(0);". It still counts as scalar context for functions that
263 care whether or not they're being called in list context.
264 User-defined subroutines may choose to care whether they are being
266 called in a void, scalar, or list context. Most subroutines do not
267 need to bother, though. That's because both scalars and lists are
268 automatically interpolated into lists. See "wantarray" in perlfunc for
269 how you would dynamically discern your function's calling context.
270 Scalar values
272 All data in Perl is a scalar, an array of scalars, or a hash of
273 scalars. A scalar may contain one single value in any of three
274 different flavors: a number, a string, or a reference. In general,
275 conversion from one form to another is transparent. Although a scalar
276 may not directly hold multiple values, it may contain a reference to an
277 array or hash which in turn contains multiple values.
278 Scalars aren't necessarily one thing or another. There's no place to
280 declare a scalar variable to be of type "string", type "number", type
281 "reference", or anything else. Because of the automatic conversion of
282 scalars, operations that return scalars don't need to care (and in
283 fact, cannot care) whether their caller is looking for a string, a
284 number, or a reference. Perl is a contextually polymorphic language
285 whose scalars can be strings, numbers, or references (which includes
286 objects). Although strings and numbers are considered pretty much the
287 same thing for nearly all purposes, references are strongly-typed,
288 uncastable pointers with builtin reference-counting and destructor
289 invocation.
290 A scalar value is interpreted as FALSE in the Boolean sense if it is
292 undefined, the null string or the number 0 (or its string equivalent,
293 "0"), and TRUE if it is anything else. The Boolean context is just a
294 special kind of scalar context where no conversion to a string or a
295 number is ever performed. Negation of a true value by "!" or "not"
296 returns a special false value. When evaluated as a string it is
297 treated as "", but as a number, it is treated as 0. Most Perl
298 operators that return true or false behave this way.
299 There are actually two varieties of null strings (sometimes referred to
301 as "empty" strings), a defined one and an undefined one. The defined
302 version is just a string of length zero, such as "". The undefined
303 version is the value that indicates that there is no real value for
304 something, such as when there was an error, or at end of file, or when
305 you refer to an uninitialized variable or element of an array or hash.
306 Although in early versions of Perl, an undefined scalar could become
307 defined when first used in a place expecting a defined value, this no
308 longer happens except for rare cases of autovivification as explained
309 in perlref. You can use the defined() operator to determine whether a
310 scalar value is defined (this has no meaning on arrays or hashes), and
311 the undef() operator to produce an undefined value.
312 To find out whether a given string is a valid non-zero number, it's
314 sometimes enough to test it against both numeric 0 and also lexical "0"
315 (although this will cause noises if warnings are on). That's because
316 strings that aren't numbers count as 0, just as they do in awk:
317 if ($str == 0 && $str ne "0") {
319 warn "That doesn't look like a number";
320 }
321 That method may be best because otherwise you won't treat IEEE
323 notations like "NaN" or "Infinity" properly. At other times, you might
324 prefer to determine whether string data can be used numerically by
325 calling the POSIX::strtod() function or by inspecting your string with
326 a regular expression (as documented in perlre).
327 warn "has nondigits" if /\D/;
329 warn "not a natural number" unless /^\d+$/; # rejects -3
330 warn "not an integer" unless /^-?\d+$/; # rejects +3
331 warn "not an integer" unless /^[+-]?\d+$/;
332 warn "not a decimal number" unless /^-?\d+\.?\d*$/; # rejects .2
333 warn "not a decimal number" unless /^-?(?:\d+(?:\.\d*)?|\.\d+)$/;
334 warn "not a C float"
335 unless /^([+-]?)(?=\d|\.\d)\d*(\.\d*)?([Ee]([+-]?\d+))?$/;
336 The length of an array is a scalar value. You may find the length of
338 array @days by evaluating $#days, as in csh. However, this isn't the
339 length of the array; it's the subscript of the last element, which is a
340 different value since there is ordinarily a 0th element. Assigning to
341 $#days actually changes the length of the array. Shortening an array
342 this way destroys intervening values. Lengthening an array that was
343 previously shortened does not recover values that were in those
344 elements.
345 You can also gain some minuscule measure of efficiency by pre-extending
347 an array that is going to get big. You can also extend an array by
348 assigning to an element that is off the end of the array. You can
349 truncate an array down to nothing by assigning the null list () to it.
350 The following are equivalent:
351 @whatever = ();
353 $#whatever = -1;
354 If you evaluate an array in scalar context, it returns the length of
356 the array. (Note that this is not true of lists, which return the last
357 value, like the C comma operator, nor of built-in functions, which
358 return whatever they feel like returning.) The following is always
359 true:
360 scalar(@whatever) == $#whatever + 1;
362 Some programmers choose to use an explicit conversion so as to leave
364 nothing to doubt:
365 $element_count = scalar(@whatever);
367 If you evaluate a hash in scalar context, it returns a false value if
369 the hash is empty. If there are any key/value pairs, it returns a true
370 value. A more precise definition is version dependent.
371 Prior to Perl 5.25 the value returned was a string consisting of the
373 number of used buckets and the number of allocated buckets, separated
374 by a slash. This is pretty much useful only to find out whether Perl's
375 internal hashing algorithm is performing poorly on your data set. For
376 example, you stick 10,000 things in a hash, but evaluating %HASH in
377 scalar context reveals "1/16", which means only one out of sixteen
378 buckets has been touched, and presumably contains all 10,000 of your
379 items. This isn't supposed to happen.
380 As of Perl 5.25 the return was changed to be the count of keys in the
382 hash. If you need access to the old behavior you can use
383 "Hash::Util::bucket_ratio()" instead.
384 If a tied hash is evaluated in scalar context, the "SCALAR" method is
386 called (with a fallback to "FIRSTKEY").
387 You can preallocate space for a hash by assigning to the keys()
389 function. This rounds up the allocated buckets to the next power of
390 two:
391 keys(%users) = 1000; # allocate 1024 buckets
393 Scalar value constructors
395 Numeric literals are specified in any of the following floating point
396 or integer formats:
397 12345
399 12345.67
400 .23E-10 # a very small number
401 3.14_15_92 # a very important number
402 4_294_967_296 # underscore for legibility
403 0xff # hex
404 0xdead_beef # more hex
405 0377 # octal (only numbers, begins with 0)
406 0o12_345 # alternative octal (introduced in Perl 5.33.5)
407 0b011011 # binary
408 0x1.999ap-4 # hexadecimal floating point (the 'p' is required)
409 You are allowed to use underscores (underbars) in numeric literals
411 between digits for legibility (but not multiple underscores in a row:
412 "23__500" is not legal; "23_500" is). You could, for example, group
413 binary digits by threes (as for a Unix-style mode argument such as
414 0b110_100_100) or by fours (to represent nibbles, as in 0b1010_0110) or
415 in other groups.
416 String literals are usually delimited by either single or double
418 quotes. They work much like quotes in the standard Unix shells:
419 double-quoted string literals are subject to backslash and variable
420 substitution; single-quoted strings are not (except for "\'" and "\\").
421 The usual C-style backslash rules apply for making characters such as
422 newline, tab, etc., as well as some more exotic forms. See "Quote and
423 Quote-like Operators" in perlop for a list.
424 Hexadecimal, octal, or binary, representations in string literals (e.g.
426 '0xff') are not automatically converted to their integer
427 representation. The hex() and oct() functions make these conversions
428 for you. See "hex" in perlfunc and "oct" in perlfunc for more details.
429 Hexadecimal floating point can start just like a hexadecimal literal,
431 and it can be followed by an optional fractional hexadecimal part, but
432 it must be followed by "p", an optional sign, and a power of two. The
433 format is useful for accurately presenting floating point values,
434 avoiding conversions to or from decimal floating point, and therefore
435 avoiding possible loss in precision. Notice that while most current
436 platforms use the 64-bit IEEE 754 floating point, not all do. Another
437 potential source of (low-order) differences are the floating point
438 rounding modes, which can differ between CPUs, operating systems, and
439 compilers, and which Perl doesn't control.
440 You can also embed newlines directly in your strings, i.e., they can
442 end on a different line than they begin. This is nice, but if you
443 forget your trailing quote, the error will not be reported until Perl
444 finds another line containing the quote character, which may be much
445 further on in the script. Variable substitution inside strings is
446 limited to scalar variables, arrays, and array or hash slices. (In
447 other words, names beginning with $ or @, followed by an optional
448 bracketed expression as a subscript.) The following code segment
449 prints out "The price is $100."
450 $Price = '$100'; # not interpolated
452 print "The price is $Price.\n"; # interpolated
453 There is no double interpolation in Perl, so the $100 is left as is.
455 By default floating point numbers substituted inside strings use the
457 dot (".") as the decimal separator. If "use locale" is in effect, and
458 POSIX::setlocale() has been called, the character used for the decimal
459 separator is affected by the LC_NUMERIC locale. See perllocale and
460 POSIX.
461 Demarcated variable names using braces
463 As in some shells, you can enclose the variable name in braces as a
465 demarcator to disambiguate it from following alphanumerics and
466 underscores or other text. You must also do this when interpolating a
467 variable into a string to separate the variable name from a following
468 double-colon or an apostrophe since these would be otherwise treated as
469 a package separator:
470 $who = "Larry";
472 print PASSWD "${who}::0:0:Superuser:/:/bin/perl\n";
473 print "We use ${who}speak when ${who}'s here.\n";
474 Without the braces, Perl would have looked for a $whospeak, a $who::0,
476 and a "$who's" variable. The last two would be the $0 and the $s
477 variables in the (presumably) non-existent package "who".
478 In fact, a simple identifier within such curly braces is forced to be a
480 string, and likewise within a hash subscript. Neither need quoting. Our
481 earlier example, $days{'Feb'} can be written as $days{Feb} and the
482 quotes will be assumed automatically. But anything more complicated in
483 the subscript will be interpreted as an expression. This means for
484 example that "$version{2.0}++" is equivalent to "$version{2}++", not to
485 "$version{'2.0'}++".
486 There is a similar problem with interpolation with text that looks like
488 array or hash access notation. Placing a simple variable like $who
489 immediately in front of text like "[1]" or "{foo}" would cause the
490 variable to be interpolated as accessing an element of @who or a value
491 stored in %who:
492 $who = "Larry Wall";
494 print "$who[1] is the father of Perl.\n";
495 would attempt to access index 1 of an array named @who. Again, using
497 braces will prevent this from happening:
498 $who = "Larry Wall";
500 print "${who}[1] is the father of Perl.\n";
501 will be treated the same as
503 $who = "Larry Wall";
505 print $who . "[1] is the father of Perl.\n";
506 This notation also applies to more complex variable descriptions, such
508 as array or hash access with subscripts. For instance
509 @name = qw(Larry Curly Moe);
511 print "Also ${name[0]}[1] was a member\n";
512 Without the braces the above example would be parsed as a two level
514 array subscript in the @name array, and under "use strict" would likely
515 produce a fatal exception, as it would be parsed like this:
516 print "Also " . $name[0][1] . " was a member\n";
518 and not as the intended:
520 print "Also " . $name[0] . "[1] was a member\n";
522 A similar result may be derived by using a backslash on the first
524 character of the subscript or package notation that is not part of the
525 variable you want to access. Thus the above example could also be
526 written:
527 @name = qw(Larry Curly Moe);
529 print "Also $name[0]\[1] was a member\n";
530 however for some special variables (multi character caret variables)
532 the demarcated form using curly braces is the only way you can
533 reference the variable at all, and the only way you can access a
534 subscript of the variable via interpolation.
535 Consider the magic array "@{^CAPTURE}" which is populated by the regex
537 engine with the contents of all of the capture buffers in a pattern
538 (see perlvar and perlre). The only way you can access one of these
539 members inside of a string is via the braced (demarcated) form:
540 "abc"=~/(.)(.)(.)/
542 and print "Second buffer is ${^CAPTURE[1]}";
543 is equivalent to
545 "abc"=~/(.)(.)(.)/
547 and print "Second buffer is " . ${^CAPTURE}[1];
548 Saying "@^CAPTURE" is a syntax error, so it must be referenced as
550 "@{^CAPTURE}", and to access one of its elements in normal code you
551 would write " ${^CAPTURE}[1] ". However when interpolating in a string
552 "${^CAPTURE}[1]" would be equivalent to "${^CAPTURE} . "[1]"", which
553 does not even refer to the same variable! Thus the subscripts must also
554 be placed inside of the braces: "${^CAPTURE[1]}".
555 The demarcated form using curly braces can be used with all the
557 different types of variable access, including array and hash slices.
558 For instance code like the following:
559 @name = qw(Larry Curly Moe);
561 local $" = " and ";
562 print "My favorites were @{name[1,2]}.\n";
563 would output
565 My favorites were Curly and Moe.
567 Special floating point: infinity (Inf) and not-a-number (NaN)
569 Floating point values include the special values "Inf" and "NaN", for
571 infinity and not-a-number. The infinity can be also negative.
572 The infinity is the result of certain math operations that overflow the
574 floating point range, like 9**9**9. The not-a-number is the result
575 when the result is undefined or unrepresentable. Though note that you
576 cannot get "NaN" from some common "undefined" or "out-of-range"
577 operations like dividing by zero, or square root of a negative number,
578 since Perl generates fatal errors for those.
579 The infinity and not-a-number have their own special arithmetic rules.
581 The general rule is that they are "contagious": "Inf" plus one is
582 "Inf", and "NaN" plus one is "NaN". Where things get interesting is
583 when you combine infinities and not-a-numbers: "Inf" minus "Inf" and
584 "Inf" divided by "Inf" are "NaN" (while "Inf" plus "Inf" is "Inf" and
585 "Inf" times "Inf" is "Inf"). "NaN" is also curious in that it does not
586 equal any number, including itself: "NaN" != "NaN".
587 Perl doesn't understand "Inf" and "NaN" as numeric literals, but you
589 can have them as strings, and Perl will convert them as needed: "Inf" +
590 1. (You can, however, import them from the POSIX extension; "use POSIX
591 qw(Inf NaN);" and then use them as literals.)
592 Note that on input (string to number) Perl accepts "Inf" and "NaN" in
594 many forms. Case is ignored, and the Win32-specific forms like
595 "1.#INF" are understood, but on output the values are normalized to
596 "Inf" and "NaN".
597 Version Strings
599 A literal of the form "v1.20.300.4000" is parsed as a string composed
601 of characters with the specified ordinals. This form, known as
602 v-strings, provides an alternative, more readable way to construct
603 strings, rather than use the somewhat less readable interpolation form
604 "\x{1}\x{14}\x{12c}\x{fa0}". This is useful for representing Unicode
605 strings, and for comparing version "numbers" using the string
606 comparison operators, "cmp", "gt", "lt" etc. If there are two or more
607 dots in the literal, the leading "v" may be omitted.
608 print v9786; # prints SMILEY, "\x{263a}"
610 print v102.111.111; # prints "foo"
611 print 102.111.111; # same
612 Such literals are accepted by both "require" and "use" for doing a
614 version check. Note that using the v-strings for IPv4 addresses is not
615 portable unless you also use the inet_aton()/inet_ntoa() routines of
616 the Socket package.
617 Note that since Perl 5.8.1 the single-number v-strings (like "v65") are
619 not v-strings before the "=>" operator (which is usually used to
620 separate a hash key from a hash value); instead they are interpreted as
621 literal strings ('v65'). They were v-strings from Perl 5.6.0 to Perl
622 5.8.0, but that caused more confusion and breakage than good. Multi-
623 number v-strings like "v65.66" and 65.66.67 continue to be v-strings
624 always.
625 Special Literals
627 The special literals __FILE__, __LINE__, and __PACKAGE__ represent the
629 current filename, line number, and package name at that point in your
630 program. __SUB__ gives a reference to the current subroutine. They
631 may be used only as separate tokens; they will not be interpolated into
632 strings. If there is no current package (due to an empty "package;"
633 directive), __PACKAGE__ is the undefined value. (But the empty
634 "package;" is no longer supported, as of version 5.10.) Outside of a
635 subroutine, __SUB__ is the undefined value. __SUB__ is only available
636 in 5.16 or higher, and only with a "use v5.16" or "use feature
637 "current_sub"" declaration.
638 The two control characters ^D and ^Z, and the tokens __END__ and
640 __DATA__ may be used to indicate the logical end of the script before
641 the actual end of file. Any following text is ignored by the
642 interpreter unless read by the program as described below.
643 Text after __DATA__ may be read via the filehandle "PACKNAME::DATA",
645 where "PACKNAME" is the package that was current when the __DATA__
646 token was encountered. The filehandle is left open pointing to the
647 line after __DATA__. The program should "close DATA" when it is done
648 reading from it. (Leaving it open leaks filehandles if the module is
649 reloaded for any reason, so it's a safer practice to close it.) For
650 compatibility with older scripts written before __DATA__ was
651 introduced, __END__ behaves like __DATA__ in the top level script (but
652 not in files loaded with "require" or "do") and leaves the remaining
653 contents of the file accessible via "main::DATA".
654 while (my $line = <DATA>) { print $line; }
656 close DATA;
657 __DATA__
658 Hello world.
659 The "DATA" file handle by default has whatever PerlIO layers were in
661 place when Perl read the file to parse the source. Normally that means
662 that the file is being read bytewise, as if it were encoded in Latin-1,
663 but there are two major ways for it to be otherwise. Firstly, if the
664 "__END__"/"__DATA__" token is in the scope of a "use utf8" pragma then
665 the "DATA" handle will be in UTF-8 mode. And secondly, if the source
666 is being read from perl's standard input then the "DATA" file handle is
667 actually aliased to the "STDIN" file handle, and may be in UTF-8 mode
668 because of the "PERL_UNICODE" environment variable or perl's command-
669 line switches.
670 See SelfLoader for more description of __DATA__, and an example of its
672 use. Note that you cannot read from the DATA filehandle in a BEGIN
673 block: the BEGIN block is executed as soon as it is seen (during
674 compilation), at which point the corresponding __DATA__ (or __END__)
675 token has not yet been seen.
676 Barewords
678 A word that has no other interpretation in the grammar will be treated
680 as if it were a quoted string. These are known as "barewords". As
681 with filehandles and labels, a bareword that consists entirely of
682 lowercase letters risks conflict with future reserved words, and if you
683 use the "use warnings" pragma or the -w switch, Perl will warn you
684 about any such words. Perl limits barewords (like identifiers) to
685 about 250 characters. Future versions of Perl are likely to eliminate
686 these arbitrary limitations.
687 Some people may wish to outlaw barewords entirely. If you say
689 use strict 'subs';
691 then any bareword that would NOT be interpreted as a subroutine call
693 produces a compile-time error instead. The restriction lasts to the
694 end of the enclosing block. An inner block may countermand this by
695 saying "no strict 'subs'".
696 Array Interpolation
698 Arrays and slices are interpolated into double-quoted strings by
700 joining the elements with the delimiter specified in the $" variable
701 ($LIST_SEPARATOR if "use English;" is specified), space by default.
702 The following are equivalent:
703 $temp = join($", @ARGV);
705 system "echo $temp";
706 system "echo @ARGV";
708 Within search patterns (which also undergo double-quotish substitution)
710 there is an unfortunate ambiguity: Is "/$foo[bar]/" to be interpreted
711 as "/${foo}[bar]/" (where "[bar]" is a character class for the regular
712 expression) or as "/${foo[bar]}/" (where "[bar]" is the subscript to
713 array @foo)? If @foo doesn't otherwise exist, then it's obviously a
714 character class. If @foo exists, Perl takes a good guess about
715 "[bar]", and is almost always right. If it does guess wrong, or if
716 you're just plain paranoid, you can force the correct interpretation
717 with curly braces as above.
718 If you're looking for the information on how to use here-documents,
720 which used to be here, that's been moved to "Quote and Quote-like
721 Operators" in perlop.
722 List value constructors
724 List values are denoted by separating individual values by commas (and
725 enclosing the list in parentheses where precedence requires it):
726 (LIST)
728 In a context not requiring a list value, the value of what appears to
730 be a list literal is simply the value of the final element, as with the
731 C comma operator. For example,
732 @foo = ('cc', '-E', $bar);
734 assigns the entire list value to array @foo, but
736 $foo = ('cc', '-E', $bar);
738 assigns the value of variable $bar to the scalar variable $foo. Note
740 that the value of an actual array in scalar context is the length of
741 the array; the following assigns the value 3 to $foo:
742 @foo = ('cc', '-E', $bar);
744 $foo = @foo; # $foo gets 3
745 You may have an optional comma before the closing parenthesis of a list
747 literal, so that you can say:
748 @foo = (
750 1,
751 2,
752 3,
753 );
754 To use a here-document to assign an array, one line per element, you
756 might use an approach like this:
757 @sauces = <<End_Lines =~ m/(\S.*\S)/g;
759 normal tomato
760 spicy tomato
761 green chile
762 pesto
763 white wine
764 End_Lines
765 LISTs do automatic interpolation of sublists. That is, when a LIST is
767 evaluated, each element of the list is evaluated in list context, and
768 the resulting list value is interpolated into LIST just as if each
769 individual element were a member of LIST. Thus arrays and hashes lose
770 their identity in a LIST--the list
771 (@foo,@bar,&SomeSub,%glarch)
773 contains all the elements of @foo followed by all the elements of @bar,
775 followed by all the elements returned by the subroutine named SomeSub
776 called in list context, followed by the key/value pairs of %glarch. To
777 make a list reference that does NOT interpolate, see perlref.
778 The null list is represented by (). Interpolating it in a list has no
780 effect. Thus ((),(),()) is equivalent to (). Similarly, interpolating
781 an array with no elements is the same as if no array had been
782 interpolated at that point.
783 This interpolation combines with the facts that the opening and closing
785 parentheses are optional (except when necessary for precedence) and
786 lists may end with an optional comma to mean that multiple commas
787 within lists are legal syntax. The list "1,,3" is a concatenation of
788 two lists, "1," and 3, the first of which ends with that optional
789 comma. "1,,3" is "(1,),(3)" is "1,3" (And similarly for "1,,,3" is
790 "(1,),(,),3" is "1,3" and so on.) Not that we'd advise you to use this
791 obfuscation.
792 A list value may also be subscripted like a normal array. You must put
794 the list in parentheses to avoid ambiguity. For example:
795 # Stat returns list value.
797 $time = (stat($file))[8];
798 # SYNTAX ERROR HERE.
800 $time = stat($file)[8]; # OOPS, FORGOT PARENTHESES
801 # Find a hex digit.
803 $hexdigit = ('a','b','c','d','e','f')[$digit-10];
804 # A "reverse comma operator".
806 return (pop(@foo),pop(@foo))[0];
807 Lists may be assigned to only when each element of the list is itself
809 legal to assign to:
810 ($x, $y, $z) = (1, 2, 3);
812 ($map{'red'}, $map{'blue'}, $map{'green'}) = (0x00f, 0x0f0, 0xf00);
814 An exception to this is that you may assign to "undef" in a list. This
816 is useful for throwing away some of the return values of a function:
817 ($dev, $ino, undef, undef, $uid, $gid) = stat($file);
819 As of Perl 5.22, you can also use "(undef)x2" instead of "undef,
821 undef". (You can also do "($x) x 2", which is less useful, because it
822 assigns to the same variable twice, clobbering the first value
823 assigned.)
824 When you assign a list of scalars to an array, all previous values in
826 that array are wiped out and the number of elements in the array will
827 now be equal to the number of elements in the right-hand list -- the
828 list from which assignment was made. The array will automatically
829 resize itself to precisely accommodate each element in the right-hand
830 list.
831 use warnings;
833 my (@xyz, $x, $y, $z);
834 @xyz = (1, 2, 3);
836 print "@xyz\n"; # 1 2 3
837 @xyz = ('al', 'be', 'ga', 'de');
839 print "@xyz\n"; # al be ga de
840 @xyz = (101, 102);
842 print "@xyz\n"; # 101 102
843 When, however, you assign a list of scalars to another list of scalars,
845 the results differ according to whether the left-hand list -- the list
846 being assigned to -- has the same, more or fewer elements than the
847 right-hand list.
848 ($x, $y, $z) = (1, 2, 3);
850 print "$x $y $z\n"; # 1 2 3
851 ($x, $y, $z) = ('al', 'be', 'ga', 'de');
853 print "$x $y $z\n"; # al be ga
854 ($x, $y, $z) = (101, 102);
856 print "$x $y $z\n"; # 101 102
857 # Use of uninitialized value $z in concatenation (.)
858 # or string at [program] line [line number].
859 If the number of scalars in the left-hand list is less than that in the
861 right-hand list, the "extra" scalars in the right-hand list will simply
862 not be assigned.
863 If the number of scalars in the left-hand list is greater than that in
865 the left-hand list, the "missing" scalars will become undefined.
866 ($x, $y, $z) = (101, 102);
868 for my $el ($x, $y, $z) {
869 (defined $el) ? print "$el " : print "<undef>";
870 }
871 print "\n";
872 # 101 102 <undef>
873 List assignment in scalar context returns the number of elements
875 produced by the expression on the right side of the assignment:
876 $x = (($foo,$bar) = (3,2,1)); # set $x to 3, not 2
878 $x = (($foo,$bar) = f()); # set $x to f()'s return count
879 This is handy when you want to do a list assignment in a Boolean
881 context, because most list functions return a null list when finished,
882 which when assigned produces a 0, which is interpreted as FALSE.
883 It's also the source of a useful idiom for executing a function or
885 performing an operation in list context and then counting the number of
886 return values, by assigning to an empty list and then using that
887 assignment in scalar context. For example, this code:
888 $count = () = $string =~ /\d+/g;
890 will place into $count the number of digit groups found in $string.
892 This happens because the pattern match is in list context (since it is
893 being assigned to the empty list), and will therefore return a list of
894 all matching parts of the string. The list assignment in scalar
895 context will translate that into the number of elements (here, the
896 number of times the pattern matched) and assign that to $count. Note
897 that simply using
898 $count = $string =~ /\d+/g;
900 would not have worked, since a pattern match in scalar context will
902 only return true or false, rather than a count of matches.
903 The final element of a list assignment may be an array or a hash:
905 ($x, $y, @rest) = split;
907 my($x, $y, %rest) = @_;
908 You can actually put an array or hash anywhere in the list, but the
910 first one in the list will soak up all the values, and anything after
911 it will become undefined. This may be useful in a my() or local().
912 A hash can be initialized using a literal list holding pairs of items
914 to be interpreted as a key and a value:
915 # same as map assignment above
917 %map = ('red',0x00f,'blue',0x0f0,'green',0xf00);
918 While literal lists and named arrays are often interchangeable, that's
920 not the case for hashes. Just because you can subscript a list value
921 like a normal array does not mean that you can subscript a list value
922 as a hash. Likewise, hashes included as parts of other lists
923 (including parameters lists and return lists from functions) always
924 flatten out into key/value pairs. That's why it's good to use
925 references sometimes.
926 It is often more readable to use the "=>" operator between key/value
928 pairs. The "=>" operator is mostly just a more visually distinctive
929 synonym for a comma, but it also arranges for its left-hand operand to
930 be interpreted as a string if it's a bareword that would be a legal
931 simple identifier. "=>" doesn't quote compound identifiers, that
932 contain double colons. This makes it nice for initializing hashes:
933 %map = (
935 red => 0x00f,
936 blue => 0x0f0,
937 green => 0xf00,
938 );
939 or for initializing hash references to be used as records:
941 $rec = {
943 witch => 'Mable the Merciless',
944 cat => 'Fluffy the Ferocious',
945 date => '10/31/1776',
946 };
947 or for using call-by-named-parameter to complicated functions:
949 $field = $query->radio_group(
951 name => 'group_name',
952 values => ['eenie','meenie','minie'],
953 default => 'meenie',
954 linebreak => 'true',
955 labels => \%labels
956 );
957 Note that just because a hash is initialized in that order doesn't mean
959 that it comes out in that order. See "sort" in perlfunc for examples
960 of how to arrange for an output ordering.
961 If a key appears more than once in the initializer list of a hash, the
963 last occurrence wins:
964 %circle = (
966 center => [5, 10],
967 center => [27, 9],
968 radius => 100,
969 color => [0xDF, 0xFF, 0x00],
970 radius => 54,
971 );
972 # same as
974 %circle = (
975 center => [27, 9],
976 color => [0xDF, 0xFF, 0x00],
977 radius => 54,
978 );
979 This can be used to provide overridable configuration defaults:
981 # values in %args take priority over %config_defaults
983 %config = (%config_defaults, %args);
984 Subscripts
986 An array can be accessed one scalar at a time by specifying a dollar
987 sign ("$"), then the name of the array (without the leading "@"), then
988 the subscript inside square brackets. For example:
989 @myarray = (5, 50, 500, 5000);
991 print "The Third Element is", $myarray[2], "\n";
992 The array indices start with 0. A negative subscript retrieves its
994 value from the end. In our example, $myarray[-1] would have been 5000,
995 and $myarray[-2] would have been 500.
996 Hash subscripts are similar, only instead of square brackets curly
998 brackets are used. For example:
999 %scientists =
1001 (
1002 "Newton" => "Isaac",
1003 "Einstein" => "Albert",
1004 "Darwin" => "Charles",
1005 "Feynman" => "Richard",
1006 );
1007 print "Darwin's First Name is ", $scientists{"Darwin"}, "\n";
1009 You can also subscript a list to get a single element from it:
1011 $dir = (getpwnam("daemon"))[7];
1013 Multi-dimensional array emulation
1015 Multidimensional arrays may be emulated by subscripting a hash with a
1016 list. The elements of the list are joined with the subscript separator
1017 (see "$;" in perlvar).
1018 $foo{$x,$y,$z}
1020 is equivalent to
1022 $foo{join($;, $x, $y, $z)}
1024 The default subscript separator is "\034", the same as SUBSEP in awk.
1026 Slices
1028 A slice accesses several elements of a list, an array, or a hash
1029 simultaneously using a list of subscripts. It's more convenient than
1030 writing out the individual elements as a list of separate scalar
1031 values.
1032 ($him, $her) = @folks[0,-1]; # array slice
1034 @them = @folks[0 .. 3]; # array slice
1035 ($who, $home) = @ENV{"USER", "HOME"}; # hash slice
1036 ($uid, $dir) = (getpwnam("daemon"))[2,7]; # list slice
1037 Since you can assign to a list of variables, you can also assign to an
1039 array or hash slice.
1040 @days[3..5] = qw/Wed Thu Fri/;
1042 @colors{'red','blue','green'}
1043 = (0xff0000, 0x0000ff, 0x00ff00);
1044 @folks[0, -1] = @folks[-1, 0];
1045 The previous assignments are exactly equivalent to
1047 ($days[3], $days[4], $days[5]) = qw/Wed Thu Fri/;
1049 ($colors{'red'}, $colors{'blue'}, $colors{'green'})
1050 = (0xff0000, 0x0000ff, 0x00ff00);
1051 ($folks[0], $folks[-1]) = ($folks[-1], $folks[0]);
1052 Since changing a slice changes the original array or hash that it's
1054 slicing, a "foreach" construct will alter some--or even all--of the
1055 values of the array or hash.
1056 foreach (@array[ 4 .. 10 ]) { s/peter/paul/ }
1058 foreach (@hash{qw[key1 key2]}) {
1060 s/^\s+//; # trim leading whitespace
1061 s/\s+$//; # trim trailing whitespace
1062 s/\b(\w)(\w*)\b/\u$1\L$2/g; # "titlecase" words
1063 }
1064 As a special exception, when you slice a list (but not an array or a
1066 hash), if the list evaluates to empty, then taking a slice of that
1067 empty list will always yield the empty list in turn. Thus:
1068 @a = ()[0,1]; # @a has no elements
1070 @b = (@a)[0,1]; # @b has no elements
1071 @c = (sub{}->())[0,1]; # @c has no elements
1072 @d = ('a','b')[0,1]; # @d has two elements
1073 @e = (@d)[0,1,8,9]; # @e has four elements
1074 @f = (@d)[8,9]; # @f has two elements
1075 This makes it easy to write loops that terminate when a null list is
1077 returned:
1078 while ( ($home, $user) = (getpwent)[7,0] ) {
1080 printf "%-8s %s\n", $user, $home;
1081 }
1082 As noted earlier in this document, the scalar sense of list assignment
1084 is the number of elements on the right-hand side of the assignment.
1085 The null list contains no elements, so when the password file is
1086 exhausted, the result is 0, not 2.
1087 Slices in scalar context return the last item of the slice.
1089 @a = qw/first second third/;
1091 %h = (first => 'A', second => 'B');
1092 $t = @a[0, 1]; # $t is now 'second'
1093 $u = @h{'first', 'second'}; # $u is now 'B'
1094 If you're confused about why you use an '@' there on a hash slice
1096 instead of a '%', think of it like this. The type of bracket (square
1097 or curly) governs whether it's an array or a hash being looked at. On
1098 the other hand, the leading symbol ('$' or '@') on the array or hash
1099 indicates whether you are getting back a singular value (a scalar) or a
1100 plural one (a list).
1101 Key/Value Hash Slices
1103 Starting in Perl 5.20, a hash slice operation with the % symbol is a
1105 variant of slice operation returning a list of key/value pairs rather
1106 than just values:
1107 %h = (blonk => 2, foo => 3, squink => 5, bar => 8);
1109 %subset = %h{'foo', 'bar'}; # key/value hash slice
1110 # %subset is now (foo => 3, bar => 8)
1111 %removed = delete %h{'foo', 'bar'};
1112 # %removed is now (foo => 3, bar => 8)
1113 # %h is now (blonk => 2, squink => 5)
1114 However, the result of such a slice cannot be localized or assigned to.
1116 These are otherwise very much consistent with hash slices using the @
1117 symbol.
1118 Index/Value Array Slices
1120 Similar to key/value hash slices (and also introduced in Perl 5.20),
1122 the % array slice syntax returns a list of index/value pairs:
1123 @a = "a".."z";
1125 @list = %a[3,4,6];
1126 # @list is now (3, "d", 4, "e", 6, "g")
1127 @removed = delete %a[3,4,6]
1128 # @removed is now (3, "d", 4, "e", 6, "g")
1129 # @list[3,4,6] are now undef
1130 Note that calling "delete" on array values is strongly discouraged.
1132 Typeglobs and Filehandles
1134 Perl uses an internal type called a typeglob to hold an entire symbol
1135 table entry. The type prefix of a typeglob is a "*", because it
1136 represents all types. This used to be the preferred way to pass arrays
1137 and hashes by reference into a function, but now that we have real
1138 references, this is seldom needed.
1139 The main use of typeglobs in modern Perl is create symbol table
1141 aliases. This assignment:
1142 *this = *that;
1144 makes $this an alias for $that, @this an alias for @that, %this an
1146 alias for %that, &this an alias for &that, etc. Much safer is to use a
1147 reference. This:
1148 local *Here::blue = \$There::green;
1150 temporarily makes $Here::blue an alias for $There::green, but doesn't
1152 make @Here::blue an alias for @There::green, or %Here::blue an alias
1153 for %There::green, etc. See "Symbol Tables" in perlmod for more
1154 examples of this. Strange though this may seem, this is the basis for
1155 the whole module import/export system.
1156 Another use for typeglobs is to pass filehandles into a function or to
1158 create new filehandles. If you need to use a typeglob to save away a
1159 filehandle, do it this way:
1160 $fh = *STDOUT;
1162 or perhaps as a real reference, like this:
1164 $fh = \*STDOUT;
1166 See perlsub for examples of using these as indirect filehandles in
1168 functions.
1169 Typeglobs are also a way to create a local filehandle using the local()
1171 operator. These last until their block is exited, but may be passed
1172 back. For example:
1173 sub newopen {
1175 my $path = shift;
1176 local *FH; # not my!
1177 open (FH, $path) or return undef;
1178 return *FH;
1179 }
1180 $fh = newopen('/etc/passwd');
1181 Now that we have the *foo{THING} notation, typeglobs aren't used as
1183 much for filehandle manipulations, although they're still needed to
1184 pass brand new file and directory handles into or out of functions.
1185 That's because *HANDLE{IO} only works if HANDLE has already been used
1186 as a handle. In other words, *FH must be used to create new symbol
1187 table entries; *foo{THING} cannot. When in doubt, use *FH.
1188 All functions that are capable of creating filehandles (open(),
1190 opendir(), pipe(), socketpair(), sysopen(), socket(), and accept())
1191 automatically create an anonymous filehandle if the handle passed to
1192 them is an uninitialized scalar variable. This allows the constructs
1193 such as "open(my $fh, ...)" and "open(local $fh,...)" to be used to
1194 create filehandles that will conveniently be closed automatically when
1195 the scope ends, provided there are no other references to them. This
1196 largely eliminates the need for typeglobs when opening filehandles that
1197 must be passed around, as in the following example:
1198 sub myopen {
1200 open my $fh, "@_"
1201 or die "Can't open '@_': $!";
1202 return $fh;
1203 }
1204 {
1206 my $f = myopen("</etc/motd");
1207 print <$f>;
1208 # $f implicitly closed here
1209 }
1210 Note that if an initialized scalar variable is used instead the result
1212 is different: "my $fh='zzz'; open($fh, ...)" is equivalent to "open(
1213 *{'zzz'}, ...)". "use strict 'refs'" forbids such practice.
1214 Another way to create anonymous filehandles is with the Symbol module
1216 or with the IO::Handle module and its ilk. These modules have the
1217 advantage of not hiding different types of the same name during the
1218 local(). See the bottom of "open" in perlfunc for an example.
1219
1221 See perlvar for a description of Perl's built-in variables and a
1222 discussion of legal variable names. See perlref, perlsub, and "Symbol
1223 Tables" in perlmod for more discussion on typeglobs and the *foo{THING}
1224 syntax.
1225perl v5.36.0 2022-08-30 PERLDATA(1)