4 perlop - Perl operators and precedence
8 In Perl, the operator determines what operation is performed,
9 independent of the type of the operands. For example C<$x + $y>
10 is always a numeric addition, and if C<$x> or C<$y> do not contain
11 numbers, an attempt is made to convert them to numbers first.
13 This is in contrast to many other dynamic languages, where the
14 operation is determined by the type of the first argument. It also
15 means that Perl has two versions of some operators, one for numeric
16 and one for string comparison. For example C<$x == $y> compares
17 two numbers for equality, and C<$x eq $y> compares two strings.
19 There are a few exceptions though: C<x> can be either string
20 repetition or list repetition, depending on the type of the left
21 operand, and C<&>, C<|> and C<^> can be either string or numeric bit
24 =head2 Operator Precedence and Associativity
25 X<operator, precedence> X<precedence> X<associativity>
27 Operator precedence and associativity work in Perl more or less like
28 they do in mathematics.
30 I<Operator precedence> means some operators are evaluated before
31 others. For example, in C<2 + 4 * 5>, the multiplication has higher
32 precedence so C<4 * 5> is evaluated first yielding C<2 + 20 ==
33 22> and not C<6 * 5 == 30>.
35 I<Operator associativity> defines what happens if a sequence of the
36 same operators is used one after another: whether the evaluator will
37 evaluate the left operations first or the right. For example, in C<8
38 - 4 - 2>, subtraction is left associative so Perl evaluates the
39 expression left to right. C<8 - 4> is evaluated first making the
40 expression C<4 - 2 == 2> and not C<8 - 2 == 6>.
42 Perl operators have the following associativity and precedence,
43 listed from highest precedence to lowest. Operators borrowed from
44 C keep the same precedence relationship with each other, even where
45 C's precedence is slightly screwy. (This makes learning Perl easier
46 for C folks.) With very few exceptions, these all operate on scalar
47 values only, not array values.
49 left terms and list operators (leftward)
53 right ! ~ \ and unary + and -
58 nonassoc named unary operators
59 nonassoc < > <= >= lt gt le ge
60 nonassoc == != <=> eq ne cmp ~~
67 right = += -= *= etc. goto last next redo dump
69 nonassoc list operators (rightward)
74 In the following sections, these operators are covered in precedence order.
76 Many operators can be overloaded for objects. See L<overload>.
78 =head2 Terms and List Operators (Leftward)
79 X<list operator> X<operator, list> X<term>
81 A TERM has the highest precedence in Perl. They include variables,
82 quote and quote-like operators, any expression in parentheses,
83 and any function whose arguments are parenthesized. Actually, there
84 aren't really functions in this sense, just list operators and unary
85 operators behaving as functions because you put parentheses around
86 the arguments. These are all documented in L<perlfunc>.
88 If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
89 is followed by a left parenthesis as the next token, the operator and
90 arguments within parentheses are taken to be of highest precedence,
91 just like a normal function call.
93 In the absence of parentheses, the precedence of list operators such as
94 C<print>, C<sort>, or C<chmod> is either very high or very low depending on
95 whether you are looking at the left side or the right side of the operator.
98 @ary = (1, 3, sort 4, 2);
99 print @ary; # prints 1324
101 the commas on the right of the sort are evaluated before the sort,
102 but the commas on the left are evaluated after. In other words,
103 list operators tend to gobble up all arguments that follow, and
104 then act like a simple TERM with regard to the preceding expression.
105 Be careful with parentheses:
107 # These evaluate exit before doing the print:
108 print($foo, exit); # Obviously not what you want.
109 print $foo, exit; # Nor is this.
111 # These do the print before evaluating exit:
112 (print $foo), exit; # This is what you want.
113 print($foo), exit; # Or this.
114 print ($foo), exit; # Or even this.
118 print ($foo & 255) + 1, "\n";
120 probably doesn't do what you expect at first glance. The parentheses
121 enclose the argument list for C<print> which is evaluated (printing
122 the result of C<$foo & 255>). Then one is added to the return value
123 of C<print> (usually 1). The result is something like this:
125 1 + 1, "\n"; # Obviously not what you meant.
127 To do what you meant properly, you must write:
129 print(($foo & 255) + 1, "\n");
131 See L<Named Unary Operators> for more discussion of this.
133 Also parsed as terms are the C<do {}> and C<eval {}> constructs, as
134 well as subroutine and method calls, and the anonymous
135 constructors C<[]> and C<{}>.
137 See also L<Quote and Quote-like Operators> toward the end of this section,
138 as well as L</"I/O Operators">.
140 =head2 The Arrow Operator
141 X<arrow> X<dereference> X<< -> >>
143 "C<< -> >>" is an infix dereference operator, just as it is in C
144 and C++. If the right side is either a C<[...]>, C<{...}>, or a
145 C<(...)> subscript, then the left side must be either a hard or
146 symbolic reference to an array, a hash, or a subroutine respectively.
147 (Or technically speaking, a location capable of holding a hard
148 reference, if it's an array or hash reference being used for
149 assignment.) See L<perlreftut> and L<perlref>.
151 Otherwise, the right side is a method name or a simple scalar
152 variable containing either the method name or a subroutine reference,
153 and the left side must be either an object (a blessed reference)
154 or a class name (that is, a package name). See L<perlobj>.
156 The dereferencing cases (as opposed to method-calling cases) are
157 somewhat extended by the experimental C<postderef> feature. For the
158 details of that feature, consult L<perlref/Postfix Dereference Syntax>.
160 =head2 Auto-increment and Auto-decrement
161 X<increment> X<auto-increment> X<++> X<decrement> X<auto-decrement> X<-->
163 "++" and "--" work as in C. That is, if placed before a variable,
164 they increment or decrement the variable by one before returning the
165 value, and if placed after, increment or decrement after returning the
169 print $i++; # prints 0
170 print ++$j; # prints 1
172 Note that just as in C, Perl doesn't define B<when> the variable is
173 incremented or decremented. You just know it will be done sometime
174 before or after the value is returned. This also means that modifying
175 a variable twice in the same statement will lead to undefined behavior.
176 Avoid statements like:
181 Perl will not guarantee what the result of the above statements is.
183 The auto-increment operator has a little extra builtin magic to it. If
184 you increment a variable that is numeric, or that has ever been used in
185 a numeric context, you get a normal increment. If, however, the
186 variable has been used in only string contexts since it was set, and
187 has a value that is not the empty string and matches the pattern
188 C</^[a-zA-Z]*[0-9]*\z/>, the increment is done as a string, preserving each
189 character within its range, with carry:
191 print ++($foo = "99"); # prints "100"
192 print ++($foo = "a0"); # prints "a1"
193 print ++($foo = "Az"); # prints "Ba"
194 print ++($foo = "zz"); # prints "aaa"
196 C<undef> is always treated as numeric, and in particular is changed
197 to C<0> before incrementing (so that a post-increment of an undef value
198 will return C<0> rather than C<undef>).
200 The auto-decrement operator is not magical.
202 =head2 Exponentiation
203 X<**> X<exponentiation> X<power>
205 Binary "**" is the exponentiation operator. It binds even more
206 tightly than unary minus, so -2**4 is -(2**4), not (-2)**4. (This is
207 implemented using C's pow(3) function, which actually works on doubles
210 =head2 Symbolic Unary Operators
211 X<unary operator> X<operator, unary>
213 Unary "!" performs logical negation, that is, "not". See also C<not> for a lower
214 precedence version of this.
217 Unary "-" performs arithmetic negation if the operand is numeric,
218 including any string that looks like a number. If the operand is
219 an identifier, a string consisting of a minus sign concatenated
220 with the identifier is returned. Otherwise, if the string starts
221 with a plus or minus, a string starting with the opposite sign is
222 returned. One effect of these rules is that -bareword is equivalent
223 to the string "-bareword". If, however, the string begins with a
224 non-alphabetic character (excluding "+" or "-"), Perl will attempt to convert
225 the string to a numeric and the arithmetic negation is performed. If the
226 string cannot be cleanly converted to a numeric, Perl will give the warning
227 B<Argument "the string" isn't numeric in negation (-) at ...>.
228 X<-> X<negation, arithmetic>
230 Unary "~" performs bitwise negation, that is, 1's complement. For
231 example, C<0666 & ~027> is 0640. (See also L<Integer Arithmetic> and
232 L<Bitwise String Operators>.) Note that the width of the result is
233 platform-dependent: ~0 is 32 bits wide on a 32-bit platform, but 64
234 bits wide on a 64-bit platform, so if you are expecting a certain bit
235 width, remember to use the "&" operator to mask off the excess bits.
236 X<~> X<negation, binary>
238 When complementing strings, if all characters have ordinal values under
239 256, then their complements will, also. But if they do not, all
240 characters will be in either 32- or 64-bit complements, depending on your
241 architecture. So for example, C<~"\x{3B1}"> is C<"\x{FFFF_FC4E}"> on
242 32-bit machines and C<"\x{FFFF_FFFF_FFFF_FC4E}"> on 64-bit machines.
244 Unary "+" has no effect whatsoever, even on strings. It is useful
245 syntactically for separating a function name from a parenthesized expression
246 that would otherwise be interpreted as the complete list of function
247 arguments. (See examples above under L<Terms and List Operators (Leftward)>.)
250 Unary "\" creates a reference to whatever follows it. See L<perlreftut>
251 and L<perlref>. Do not confuse this behavior with the behavior of
252 backslash within a string, although both forms do convey the notion
253 of protecting the next thing from interpolation.
254 X<\> X<reference> X<backslash>
256 =head2 Binding Operators
257 X<binding> X<operator, binding> X<=~> X<!~>
259 Binary "=~" binds a scalar expression to a pattern match. Certain operations
260 search or modify the string $_ by default. This operator makes that kind
261 of operation work on some other string. The right argument is a search
262 pattern, substitution, or transliteration. The left argument is what is
263 supposed to be searched, substituted, or transliterated instead of the default
264 $_. When used in scalar context, the return value generally indicates the
265 success of the operation. The exceptions are substitution (s///)
266 and transliteration (y///) with the C</r> (non-destructive) option,
267 which cause the B<r>eturn value to be the result of the substitution.
268 Behavior in list context depends on the particular operator.
269 See L</"Regexp Quote-Like Operators"> for details and L<perlretut> for
270 examples using these operators.
272 If the right argument is an expression rather than a search pattern,
273 substitution, or transliteration, it is interpreted as a search pattern at run
274 time. Note that this means that its
275 contents will be interpolated twice, so
279 is not ok, as the regex engine will end up trying to compile the
280 pattern C<\>, which it will consider a syntax error.
282 Binary "!~" is just like "=~" except the return value is negated in
285 Binary "!~" with a non-destructive substitution (s///r) or transliteration
286 (y///r) is a syntax error.
288 =head2 Multiplicative Operators
289 X<operator, multiplicative>
291 Binary "*" multiplies two numbers.
294 Binary "/" divides two numbers.
297 Binary "%" is the modulo operator, which computes the division
298 remainder of its first argument with respect to its second argument.
300 operands C<$m> and C<$n>: If C<$n> is positive, then C<$m % $n> is
301 C<$m> minus the largest multiple of C<$n> less than or equal to
302 C<$m>. If C<$n> is negative, then C<$m % $n> is C<$m> minus the
303 smallest multiple of C<$n> that is not less than C<$m> (that is, the
304 result will be less than or equal to zero). If the operands
305 C<$m> and C<$n> are floating point values and the absolute value of
306 C<$n> (that is C<abs($n)>) is less than C<(UV_MAX + 1)>, only
307 the integer portion of C<$m> and C<$n> will be used in the operation
308 (Note: here C<UV_MAX> means the maximum of the unsigned integer type).
309 If the absolute value of the right operand (C<abs($n)>) is greater than
310 or equal to C<(UV_MAX + 1)>, "%" computes the floating-point remainder
311 C<$r> in the equation C<($r = $m - $i*$n)> where C<$i> is a certain
312 integer that makes C<$r> have the same sign as the right operand
313 C<$n> (B<not> as the left operand C<$m> like C function C<fmod()>)
314 and the absolute value less than that of C<$n>.
315 Note that when C<use integer> is in scope, "%" gives you direct access
316 to the modulo operator as implemented by your C compiler. This
317 operator is not as well defined for negative operands, but it will
319 X<%> X<remainder> X<modulo> X<mod>
321 Binary "x" is the repetition operator. In scalar context or if the left
322 operand is not enclosed in parentheses, it returns a string consisting
323 of the left operand repeated the number of times specified by the right
324 operand. In list context, if the left operand is enclosed in
325 parentheses or is a list formed by C<qw/STRING/>, it repeats the list.
326 If the right operand is zero or negative (raising a warning on
327 negative), it returns an empty string
328 or an empty list, depending on the context.
331 print '-' x 80; # print row of dashes
333 print "\t" x ($tab/8), ' ' x ($tab%8); # tab over
335 @ones = (1) x 80; # a list of 80 1's
336 @ones = (5) x @ones; # set all elements to 5
339 =head2 Additive Operators
340 X<operator, additive>
342 Binary C<+> returns the sum of two numbers.
345 Binary C<-> returns the difference of two numbers.
348 Binary C<.> concatenates two strings.
349 X<string, concatenation> X<concatenation>
350 X<cat> X<concat> X<concatenate> X<.>
352 =head2 Shift Operators
353 X<shift operator> X<operator, shift> X<<< << >>>
354 X<<< >> >>> X<right shift> X<left shift> X<bitwise shift>
355 X<shl> X<shr> X<shift, right> X<shift, left>
357 Binary C<<< << >>> returns the value of its left argument shifted left by the
358 number of bits specified by the right argument. Arguments should be
359 integers. (See also L<Integer Arithmetic>.)
361 Binary C<<< >> >>> returns the value of its left argument shifted right by
362 the number of bits specified by the right argument. Arguments should
363 be integers. (See also L<Integer Arithmetic>.)
365 Note that both C<<< << >>> and C<<< >> >>> in Perl are implemented directly using
366 C<<< << >>> and C<<< >> >>> in C. If C<use integer> (see L<Integer Arithmetic>) is
367 in force then signed C integers are used, else unsigned C integers are
368 used. Either way, the implementation isn't going to generate results
369 larger than the size of the integer type Perl was built with (32 bits
372 The result of overflowing the range of the integers is undefined
373 because it is undefined also in C. In other words, using 32-bit
374 integers, C<< 1 << 32 >> is undefined. Shifting by a negative number
375 of bits is also undefined.
377 If you get tired of being subject to your platform's native integers,
378 the C<use bigint> pragma neatly sidesteps the issue altogether:
380 print 20 << 20; # 20971520
381 print 20 << 40; # 5120 on 32-bit machines,
382 # 21990232555520 on 64-bit machines
384 print 20 << 100; # 25353012004564588029934064107520
386 =head2 Named Unary Operators
387 X<operator, named unary>
389 The various named unary operators are treated as functions with one
390 argument, with optional parentheses.
392 If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
393 is followed by a left parenthesis as the next token, the operator and
394 arguments within parentheses are taken to be of highest precedence,
395 just like a normal function call. For example,
396 because named unary operators are higher precedence than C<||>:
398 chdir $foo || die; # (chdir $foo) || die
399 chdir($foo) || die; # (chdir $foo) || die
400 chdir ($foo) || die; # (chdir $foo) || die
401 chdir +($foo) || die; # (chdir $foo) || die
403 but, because * is higher precedence than named operators:
405 chdir $foo * 20; # chdir ($foo * 20)
406 chdir($foo) * 20; # (chdir $foo) * 20
407 chdir ($foo) * 20; # (chdir $foo) * 20
408 chdir +($foo) * 20; # chdir ($foo * 20)
410 rand 10 * 20; # rand (10 * 20)
411 rand(10) * 20; # (rand 10) * 20
412 rand (10) * 20; # (rand 10) * 20
413 rand +(10) * 20; # rand (10 * 20)
415 Regarding precedence, the filetest operators, like C<-f>, C<-M>, etc. are
416 treated like named unary operators, but they don't follow this functional
417 parenthesis rule. That means, for example, that C<-f($file).".bak"> is
418 equivalent to C<-f "$file.bak">.
419 X<-X> X<filetest> X<operator, filetest>
421 See also L<"Terms and List Operators (Leftward)">.
423 =head2 Relational Operators
424 X<relational operator> X<operator, relational>
426 Perl operators that return true or false generally return values
427 that can be safely used as numbers. For example, the relational
428 operators in this section and the equality operators in the next
429 one return C<1> for true and a special version of the defined empty
430 string, C<"">, which counts as a zero but is exempt from warnings
431 about improper numeric conversions, just as C<"0 but true"> is.
433 Binary "<" returns true if the left argument is numerically less than
437 Binary ">" returns true if the left argument is numerically greater
438 than the right argument.
441 Binary "<=" returns true if the left argument is numerically less than
442 or equal to the right argument.
445 Binary ">=" returns true if the left argument is numerically greater
446 than or equal to the right argument.
449 Binary "lt" returns true if the left argument is stringwise less than
453 Binary "gt" returns true if the left argument is stringwise greater
454 than the right argument.
457 Binary "le" returns true if the left argument is stringwise less than
458 or equal to the right argument.
461 Binary "ge" returns true if the left argument is stringwise greater
462 than or equal to the right argument.
465 =head2 Equality Operators
466 X<equality> X<equal> X<equals> X<operator, equality>
468 Binary "==" returns true if the left argument is numerically equal to
472 Binary "!=" returns true if the left argument is numerically not equal
473 to the right argument.
476 Binary "<=>" returns -1, 0, or 1 depending on whether the left
477 argument is numerically less than, equal to, or greater than the right
478 argument. If your platform supports NaNs (not-a-numbers) as numeric
479 values, using them with "<=>" returns undef. NaN is not "<", "==", ">",
480 "<=" or ">=" anything (even NaN), so those 5 return false. NaN != NaN
481 returns true, as does NaN != anything else. If your platform doesn't
482 support NaNs then NaN is just a string with numeric value 0.
483 X<< <=> >> X<spaceship>
485 $ perl -le '$x = "NaN"; print "No NaN support here" if $x == $x'
486 $ perl -le '$x = "NaN"; print "NaN support here" if $x != $x'
488 (Note that the L<bigint>, L<bigrat>, and L<bignum> pragmas all
491 Binary "eq" returns true if the left argument is stringwise equal to
495 Binary "ne" returns true if the left argument is stringwise not equal
496 to the right argument.
499 Binary "cmp" returns -1, 0, or 1 depending on whether the left
500 argument is stringwise less than, equal to, or greater than the right
504 Binary "~~" does a smartmatch between its arguments. Smart matching
505 is described in the next section.
508 "lt", "le", "ge", "gt" and "cmp" use the collation (sort) order specified
509 by the current locale if a legacy C<use locale> (but not
510 C<use locale ':not_characters'>) is in effect. See
511 L<perllocale>. Do not mix these with Unicode, only with legacy binary
512 encodings. The standard L<Unicode::Collate> and
513 L<Unicode::Collate::Locale> modules offer much more powerful solutions to
516 =head2 Smartmatch Operator
518 First available in Perl 5.10.1 (the 5.10.0 version behaved differently),
519 binary C<~~> does a "smartmatch" between its arguments. This is mostly
520 used implicitly in the C<when> construct described in L<perlsyn>, although
521 not all C<when> clauses call the smartmatch operator. Unique among all of
522 Perl's operators, the smartmatch operator can recurse. The smartmatch
523 operator is L<experimental|perlpolicy/experimental> and its behavior is
526 It is also unique in that all other Perl operators impose a context
527 (usually string or numeric context) on their operands, autoconverting
528 those operands to those imposed contexts. In contrast, smartmatch
529 I<infers> contexts from the actual types of its operands and uses that
530 type information to select a suitable comparison mechanism.
532 The C<~~> operator compares its operands "polymorphically", determining how
533 to compare them according to their actual types (numeric, string, array,
534 hash, etc.) Like the equality operators with which it shares the same
535 precedence, C<~~> returns 1 for true and C<""> for false. It is often best
536 read aloud as "in", "inside of", or "is contained in", because the left
537 operand is often looked for I<inside> the right operand. That makes the
538 order of the operands to the smartmatch operand often opposite that of
539 the regular match operator. In other words, the "smaller" thing is usually
540 placed in the left operand and the larger one in the right.
542 The behavior of a smartmatch depends on what type of things its arguments
543 are, as determined by the following table. The first row of the table
544 whose types apply determines the smartmatch behavior. Because what
545 actually happens is mostly determined by the type of the second operand,
546 the table is sorted on the right operand instead of on the left.
548 Left Right Description and pseudocode
549 ===============================================================
550 Any undef check whether Any is undefined
553 Any Object invoke ~~ overloading on Object, or die
555 Right operand is an ARRAY:
557 Left Right Description and pseudocode
558 ===============================================================
559 ARRAY1 ARRAY2 recurse on paired elements of ARRAY1 and ARRAY2[2]
560 like: (ARRAY1[0] ~~ ARRAY2[0])
561 && (ARRAY1[1] ~~ ARRAY2[1]) && ...
562 HASH ARRAY any ARRAY elements exist as HASH keys
563 like: grep { exists HASH->{$_} } ARRAY
564 Regexp ARRAY any ARRAY elements pattern match Regexp
565 like: grep { /Regexp/ } ARRAY
566 undef ARRAY undef in ARRAY
567 like: grep { !defined } ARRAY
568 Any ARRAY smartmatch each ARRAY element[3]
569 like: grep { Any ~~ $_ } ARRAY
571 Right operand is a HASH:
573 Left Right Description and pseudocode
574 ===============================================================
575 HASH1 HASH2 all same keys in both HASHes
577 grep { exists HASH2->{$_} } keys HASH1
578 ARRAY HASH any ARRAY elements exist as HASH keys
579 like: grep { exists HASH->{$_} } ARRAY
580 Regexp HASH any HASH keys pattern match Regexp
581 like: grep { /Regexp/ } keys HASH
582 undef HASH always false (undef can't be a key)
584 Any HASH HASH key existence
585 like: exists HASH->{Any}
587 Right operand is CODE:
589 Left Right Description and pseudocode
590 ===============================================================
591 ARRAY CODE sub returns true on all ARRAY elements[1]
592 like: !grep { !CODE->($_) } ARRAY
593 HASH CODE sub returns true on all HASH keys[1]
594 like: !grep { !CODE->($_) } keys HASH
595 Any CODE sub passed Any returns true
598 Right operand is a Regexp:
600 Left Right Description and pseudocode
601 ===============================================================
602 ARRAY Regexp any ARRAY elements match Regexp
603 like: grep { /Regexp/ } ARRAY
604 HASH Regexp any HASH keys match Regexp
605 like: grep { /Regexp/ } keys HASH
606 Any Regexp pattern match
607 like: Any =~ /Regexp/
611 Left Right Description and pseudocode
612 ===============================================================
613 Object Any invoke ~~ overloading on Object,
616 Any Num numeric equality
618 Num nummy[4] numeric equality
620 undef Any check whether undefined
622 Any Any string equality
631 Empty hashes or arrays match.
634 That is, each element smartmatches the element of the same index in the other array.[3]
637 If a circular reference is found, fall back to referential equality.
640 Either an actual number, or a string that looks like one.
644 The smartmatch implicitly dereferences any non-blessed hash or array
645 reference, so the C<I<HASH>> and C<I<ARRAY>> entries apply in those cases.
646 For blessed references, the C<I<Object>> entries apply. Smartmatches
647 involving hashes only consider hash keys, never hash values.
649 The "like" code entry is not always an exact rendition. For example, the
650 smartmatch operator short-circuits whenever possible, but C<grep> does
651 not. Also, C<grep> in scalar context returns the number of matches, but
652 C<~~> returns only true or false.
654 Unlike most operators, the smartmatch operator knows to treat C<undef>
658 @array = (1, 2, 3, undef, 4, 5);
659 say "some elements undefined" if undef ~~ @array;
661 Each operand is considered in a modified scalar context, the modification
662 being that array and hash variables are passed by reference to the
663 operator, which implicitly dereferences them. Both elements
664 of each pair are the same:
668 my %hash = (red => 1, blue => 2, green => 3,
669 orange => 4, yellow => 5, purple => 6,
670 black => 7, grey => 8, white => 9);
672 my @array = qw(red blue green);
674 say "some array elements in hash keys" if @array ~~ %hash;
675 say "some array elements in hash keys" if \@array ~~ \%hash;
677 say "red in array" if "red" ~~ @array;
678 say "red in array" if "red" ~~ \@array;
680 say "some keys end in e" if /e$/ ~~ %hash;
681 say "some keys end in e" if /e$/ ~~ \%hash;
683 Two arrays smartmatch if each element in the first array smartmatches
684 (that is, is "in") the corresponding element in the second array,
688 my @little = qw(red blue green);
689 my @bigger = ("red", "blue", [ "orange", "green" ] );
690 if (@little ~~ @bigger) { # true!
691 say "little is contained in bigger";
694 Because the smartmatch operator recurses on nested arrays, this
695 will still report that "red" is in the array.
698 my @array = qw(red blue green);
699 my $nested_array = [[[[[[[ @array ]]]]]]];
700 say "red in array" if "red" ~~ $nested_array;
702 If two arrays smartmatch each other, then they are deep
703 copies of each others' values, as this example reports:
706 my @a = (0, 1, 2, [3, [4, 5], 6], 7);
707 my @b = (0, 1, 2, [3, [4, 5], 6], 7);
709 if (@a ~~ @b && @b ~~ @a) {
710 say "a and b are deep copies of each other";
713 say "a smartmatches in b";
716 say "b smartmatches in a";
719 say "a and b don't smartmatch each other at all";
723 If you were to set C<$b[3] = 4>, then instead of reporting that "a and b
724 are deep copies of each other", it now reports that "b smartmatches in a".
725 That because the corresponding position in C<@a> contains an array that
726 (eventually) has a 4 in it.
728 Smartmatching one hash against another reports whether both contain the
729 same keys, no more and no less. This could be used to see whether two
730 records have the same field names, without caring what values those fields
731 might have. For example:
735 state $REQUIRED_FIELDS = { name=>1, rank=>1, serial_num=>1 };
737 my ($class, $init_fields) = @_;
739 die "Must supply (only) name, rank, and serial number"
740 unless $init_fields ~~ $REQUIRED_FIELDS;
745 or, if other non-required fields are allowed, use ARRAY ~~ HASH:
749 state $REQUIRED_FIELDS = { name=>1, rank=>1, serial_num=>1 };
751 my ($class, $init_fields) = @_;
753 die "Must supply (at least) name, rank, and serial number"
754 unless [keys %{$init_fields}] ~~ $REQUIRED_FIELDS;
759 The smartmatch operator is most often used as the implicit operator of a
760 C<when> clause. See the section on "Switch Statements" in L<perlsyn>.
762 =head3 Smartmatching of Objects
764 To avoid relying on an object's underlying representation, if the
765 smartmatch's right operand is an object that doesn't overload C<~~>,
766 it raises the exception "C<Smartmatching a non-overloaded object
767 breaks encapsulation>". That's because one has no business digging
768 around to see whether something is "in" an object. These are all
769 illegal on objects without a C<~~> overload:
775 However, you can change the way an object is smartmatched by overloading
776 the C<~~> operator. This is allowed to
777 extend the usual smartmatch semantics.
778 For objects that do have an C<~~> overload, see L<overload>.
780 Using an object as the left operand is allowed, although not very useful.
781 Smartmatching rules take precedence over overloading, so even if the
782 object in the left operand has smartmatch overloading, this will be
783 ignored. A left operand that is a non-overloaded object falls back on a
784 string or numeric comparison of whatever the C<ref> operator returns. That
789 does I<not> invoke the overload method with C<I<X>> as an argument.
790 Instead the above table is consulted as normal, and based on the type of
791 C<I<X>>, overloading may or may not be invoked. For simple strings or
792 numbers, in becomes equivalent to this:
794 $object ~~ $number ref($object) == $number
795 $object ~~ $string ref($object) eq $string
797 For example, this reports that the handle smells IOish
798 (but please don't really do this!):
801 my $fh = IO::Handle->new();
802 if ($fh ~~ /\bIO\b/) {
803 say "handle smells IOish";
806 That's because it treats C<$fh> as a string like
807 C<"IO::Handle=GLOB(0x8039e0)">, then pattern matches against that.
810 X<operator, bitwise, and> X<bitwise and> X<&>
812 Binary "&" returns its operands ANDed together bit by bit. Although no
813 warning is currently raised, the result is not well defined when this operation
814 is performed on operands that aren't either numbers (see
815 L<Integer Arithmetic>) or bitstrings (see L<Bitwise String Operators>).
817 Note that "&" has lower priority than relational operators, so for example
818 the parentheses are essential in a test like
820 print "Even\n" if ($x & 1) == 0;
822 =head2 Bitwise Or and Exclusive Or
823 X<operator, bitwise, or> X<bitwise or> X<|> X<operator, bitwise, xor>
826 Binary "|" returns its operands ORed together bit by bit.
828 Binary "^" returns its operands XORed together bit by bit.
830 Although no warning is currently raised, the results are not well
831 defined when these operations are performed on operands that aren't either
832 numbers (see L<Integer Arithmetic>) or bitstrings (see L<Bitwise String
835 Note that "|" and "^" have lower priority than relational operators, so
836 for example the brackets are essential in a test like
838 print "false\n" if (8 | 2) != 10;
840 =head2 C-style Logical And
841 X<&&> X<logical and> X<operator, logical, and>
843 Binary "&&" performs a short-circuit logical AND operation. That is,
844 if the left operand is false, the right operand is not even evaluated.
845 Scalar or list context propagates down to the right operand if it
848 =head2 C-style Logical Or
849 X<||> X<operator, logical, or>
851 Binary "||" performs a short-circuit logical OR operation. That is,
852 if the left operand is true, the right operand is not even evaluated.
853 Scalar or list context propagates down to the right operand if it
856 =head2 Logical Defined-Or
857 X<//> X<operator, logical, defined-or>
859 Although it has no direct equivalent in C, Perl's C<//> operator is related
860 to its C-style or. In fact, it's exactly the same as C<||>, except that it
861 tests the left hand side's definedness instead of its truth. Thus,
862 C<< EXPR1 // EXPR2 >> returns the value of C<< EXPR1 >> if it's defined,
863 otherwise, the value of C<< EXPR2 >> is returned.
864 (C<< EXPR1 >> is evaluated in scalar context, C<< EXPR2 >>
865 in the context of C<< // >> itself). Usually,
866 this is the same result as C<< defined(EXPR1) ? EXPR1 : EXPR2 >> (except that
867 the ternary-operator form can be used as a lvalue, while C<< EXPR1 // EXPR2 >>
868 cannot). This is very useful for
869 providing default values for variables. If you actually want to test if
870 at least one of C<$x> and C<$y> is defined, use C<defined($x // $y)>.
872 The C<||>, C<//> and C<&&> operators return the last value evaluated
873 (unlike C's C<||> and C<&&>, which return 0 or 1). Thus, a reasonably
874 portable way to find out the home directory might be:
879 // die "You're homeless!\n";
881 In particular, this means that you shouldn't use this
882 for selecting between two aggregates for assignment:
884 @a = @b || @c; # this is wrong
885 @a = scalar(@b) || @c; # really meant this
886 @a = @b ? @b : @c; # this works fine, though
888 As alternatives to C<&&> and C<||> when used for
889 control flow, Perl provides the C<and> and C<or> operators (see below).
890 The short-circuit behavior is identical. The precedence of "and"
891 and "or" is much lower, however, so that you can safely use them after a
892 list operator without the need for parentheses:
894 unlink "alpha", "beta", "gamma"
895 or gripe(), next LINE;
897 With the C-style operators that would have been written like this:
899 unlink("alpha", "beta", "gamma")
900 || (gripe(), next LINE);
902 It would be even more readable to write that this way:
904 unless(unlink("alpha", "beta", "gamma")) {
909 Using "or" for assignment is unlikely to do what you want; see below.
911 =head2 Range Operators
912 X<operator, range> X<range> X<..> X<...>
914 Binary ".." is the range operator, which is really two different
915 operators depending on the context. In list context, it returns a
916 list of values counting (up by ones) from the left value to the right
917 value. If the left value is greater than the right value then it
918 returns the empty list. The range operator is useful for writing
919 C<foreach (1..10)> loops and for doing slice operations on arrays. In
920 the current implementation, no temporary array is created when the
921 range operator is used as the expression in C<foreach> loops, but older
922 versions of Perl might burn a lot of memory when you write something
925 for (1 .. 1_000_000) {
929 The range operator also works on strings, using the magical
930 auto-increment, see below.
932 In scalar context, ".." returns a boolean value. The operator is
933 bistable, like a flip-flop, and emulates the line-range (comma)
934 operator of B<sed>, B<awk>, and various editors. Each ".." operator
935 maintains its own boolean state, even across calls to a subroutine
936 that contains it. It is false as long as its left operand is false.
937 Once the left operand is true, the range operator stays true until the
938 right operand is true, I<AFTER> which the range operator becomes false
939 again. It doesn't become false till the next time the range operator
940 is evaluated. It can test the right operand and become false on the
941 same evaluation it became true (as in B<awk>), but it still returns
942 true once. If you don't want it to test the right operand until the
943 next evaluation, as in B<sed>, just use three dots ("...") instead of
944 two. In all other regards, "..." behaves just like ".." does.
946 The right operand is not evaluated while the operator is in the
947 "false" state, and the left operand is not evaluated while the
948 operator is in the "true" state. The precedence is a little lower
949 than || and &&. The value returned is either the empty string for
950 false, or a sequence number (beginning with 1) for true. The sequence
951 number is reset for each range encountered. The final sequence number
952 in a range has the string "E0" appended to it, which doesn't affect
953 its numeric value, but gives you something to search for if you want
954 to exclude the endpoint. You can exclude the beginning point by
955 waiting for the sequence number to be greater than 1.
957 If either operand of scalar ".." is a constant expression,
958 that operand is considered true if it is equal (C<==>) to the current
959 input line number (the C<$.> variable).
961 To be pedantic, the comparison is actually C<int(EXPR) == int(EXPR)>,
962 but that is only an issue if you use a floating point expression; when
963 implicitly using C<$.> as described in the previous paragraph, the
964 comparison is C<int(EXPR) == int($.)> which is only an issue when C<$.>
965 is set to a floating point value and you are not reading from a file.
966 Furthermore, C<"span" .. "spat"> or C<2.18 .. 3.14> will not do what
967 you want in scalar context because each of the operands are evaluated
968 using their integer representation.
972 As a scalar operator:
974 if (101 .. 200) { print; } # print 2nd hundred lines, short for
975 # if ($. == 101 .. $. == 200) { print; }
977 next LINE if (1 .. /^$/); # skip header lines, short for
978 # next LINE if ($. == 1 .. /^$/);
979 # (typically in a loop labeled LINE)
981 s/^/> / if (/^$/ .. eof()); # quote body
983 # parse mail messages
985 $in_header = 1 .. /^$/;
986 $in_body = /^$/ .. eof;
993 close ARGV if eof; # reset $. each file
996 Here's a simple example to illustrate the difference between
997 the two range operators:
1010 This program will print only the line containing "Bar". If
1011 the range operator is changed to C<...>, it will also print the
1014 And now some examples as a list operator:
1016 for (101 .. 200) { print } # print $_ 100 times
1017 @foo = @foo[0 .. $#foo]; # an expensive no-op
1018 @foo = @foo[$#foo-4 .. $#foo]; # slice last 5 items
1020 The range operator (in list context) makes use of the magical
1021 auto-increment algorithm if the operands are strings. You
1024 @alphabet = ("A" .. "Z");
1026 to get all normal letters of the English alphabet, or
1028 $hexdigit = (0 .. 9, "a" .. "f")[$num & 15];
1030 to get a hexadecimal digit, or
1032 @z2 = ("01" .. "31");
1035 to get dates with leading zeros.
1037 If the final value specified is not in the sequence that the magical
1038 increment would produce, the sequence goes until the next value would
1039 be longer than the final value specified.
1041 If the initial value specified isn't part of a magical increment
1042 sequence (that is, a non-empty string matching C</^[a-zA-Z]*[0-9]*\z/>),
1043 only the initial value will be returned. So the following will only
1046 use charnames "greek";
1047 my @greek_small = ("\N{alpha}" .. "\N{omega}");
1049 To get the 25 traditional lowercase Greek letters, including both sigmas,
1050 you could use this instead:
1052 use charnames "greek";
1053 my @greek_small = map { chr } ( ord("\N{alpha}")
1058 However, because there are I<many> other lowercase Greek characters than
1059 just those, to match lowercase Greek characters in a regular expression,
1060 you could use the pattern C</(?:(?=\p{Greek})\p{Lower})+/> (or the
1061 L<experimental feature|perlrecharclass/Extended Bracketed Character
1062 Classes> C<S</(?[ \p{Greek} & \p{Lower} ])+/>>).
1064 Because each operand is evaluated in integer form, C<2.18 .. 3.14> will
1065 return two elements in list context.
1067 @list = (2.18 .. 3.14); # same as @list = (2 .. 3);
1069 =head2 Conditional Operator
1070 X<operator, conditional> X<operator, ternary> X<ternary> X<?:>
1072 Ternary "?:" is the conditional operator, just as in C. It works much
1073 like an if-then-else. If the argument before the ? is true, the
1074 argument before the : is returned, otherwise the argument after the :
1075 is returned. For example:
1077 printf "I have %d dog%s.\n", $n,
1078 ($n == 1) ? "" : "s";
1080 Scalar or list context propagates downward into the 2nd
1081 or 3rd argument, whichever is selected.
1083 $x = $ok ? $y : $z; # get a scalar
1084 @x = $ok ? @y : @z; # get an array
1085 $x = $ok ? @y : @z; # oops, that's just a count!
1087 The operator may be assigned to if both the 2nd and 3rd arguments are
1088 legal lvalues (meaning that you can assign to them):
1090 ($x_or_y ? $x : $y) = $z;
1092 Because this operator produces an assignable result, using assignments
1093 without parentheses will get you in trouble. For example, this:
1095 $x % 2 ? $x += 10 : $x += 2
1099 (($x % 2) ? ($x += 10) : $x) += 2
1103 ($x % 2) ? ($x += 10) : ($x += 2)
1105 That should probably be written more simply as:
1107 $x += ($x % 2) ? 10 : 2;
1109 =head2 Assignment Operators
1110 X<assignment> X<operator, assignment> X<=> X<**=> X<+=> X<*=> X<&=>
1111 X<<< <<= >>> X<&&=> X<-=> X</=> X<|=> X<<< >>= >>> X<||=> X<//=> X<.=>
1114 "=" is the ordinary assignment operator.
1116 Assignment operators work as in C. That is,
1124 although without duplicating any side effects that dereferencing the lvalue
1125 might trigger, such as from tie(). Other assignment operators work similarly.
1126 The following are recognized:
1128 **= += *= &= <<= &&=
1133 Although these are grouped by family, they all have the precedence
1134 of assignment. These combined assignment operators can only operate on
1135 scalars, whereas the ordinary assignment operator can assign to arrays,
1136 hashes, lists and even references. (See L<"Context"|perldata/Context>
1137 and L<perldata/List value constructors>, and L<perlref/Assigning to
1140 Unlike in C, the scalar assignment operator produces a valid lvalue.
1141 Modifying an assignment is equivalent to doing the assignment and
1142 then modifying the variable that was assigned to. This is useful
1143 for modifying a copy of something, like this:
1145 ($tmp = $global) =~ tr/13579/24680/;
1147 Although as of 5.14, that can be also be accomplished this way:
1150 $tmp = ($global =~ tr/13579/24680/r);
1161 Similarly, a list assignment in list context produces the list of
1162 lvalues assigned to, and a list assignment in scalar context returns
1163 the number of elements produced by the expression on the right hand
1164 side of the assignment.
1166 =head2 Comma Operator
1167 X<comma> X<operator, comma> X<,>
1169 Binary "," is the comma operator. In scalar context it evaluates
1170 its left argument, throws that value away, then evaluates its right
1171 argument and returns that value. This is just like C's comma operator.
1173 In list context, it's just the list argument separator, and inserts
1174 both its arguments into the list. These arguments are also evaluated
1177 The C<< => >> operator is a synonym for the comma except that it causes a
1178 word on its left to be interpreted as a string if it begins with a letter
1179 or underscore and is composed only of letters, digits and underscores.
1180 This includes operands that might otherwise be interpreted as operators,
1181 constants, single number v-strings or function calls. If in doubt about
1182 this behavior, the left operand can be quoted explicitly.
1184 Otherwise, the C<< => >> operator behaves exactly as the comma operator
1185 or list argument separator, according to context.
1189 use constant FOO => "something";
1191 my %h = ( FOO => 23 );
1195 my %h = ("FOO", 23);
1199 my %h = ("something", 23);
1201 The C<< => >> operator is helpful in documenting the correspondence
1202 between keys and values in hashes, and other paired elements in lists.
1204 %hash = ( $key => $value );
1205 login( $username => $password );
1207 The special quoting behavior ignores precedence, and hence may apply to
1208 I<part> of the left operand:
1210 print time.shift => "bbb";
1212 That example prints something like "1314363215shiftbbb", because the
1213 C<< => >> implicitly quotes the C<shift> immediately on its left, ignoring
1214 the fact that C<time.shift> is the entire left operand.
1216 =head2 List Operators (Rightward)
1217 X<operator, list, rightward> X<list operator>
1219 On the right side of a list operator, the comma has very low precedence,
1220 such that it controls all comma-separated expressions found there.
1221 The only operators with lower precedence are the logical operators
1222 "and", "or", and "not", which may be used to evaluate calls to list
1223 operators without the need for parentheses:
1225 open HANDLE, "< :utf8", "filename" or die "Can't open: $!\n";
1227 However, some people find that code harder to read than writing
1228 it with parentheses:
1230 open(HANDLE, "< :utf8", "filename") or die "Can't open: $!\n";
1232 in which case you might as well just use the more customary "||" operator:
1234 open(HANDLE, "< :utf8", "filename") || die "Can't open: $!\n";
1236 See also discussion of list operators in L<Terms and List Operators (Leftward)>.
1239 X<operator, logical, not> X<not>
1241 Unary "not" returns the logical negation of the expression to its right.
1242 It's the equivalent of "!" except for the very low precedence.
1245 X<operator, logical, and> X<and>
1247 Binary "and" returns the logical conjunction of the two surrounding
1248 expressions. It's equivalent to C<&&> except for the very low
1249 precedence. This means that it short-circuits: the right
1250 expression is evaluated only if the left expression is true.
1252 =head2 Logical or and Exclusive Or
1253 X<operator, logical, or> X<operator, logical, xor>
1254 X<operator, logical, exclusive or>
1257 Binary "or" returns the logical disjunction of the two surrounding
1258 expressions. It's equivalent to C<||> except for the very low precedence.
1259 This makes it useful for control flow:
1261 print FH $data or die "Can't write to FH: $!";
1263 This means that it short-circuits: the right expression is evaluated
1264 only if the left expression is false. Due to its precedence, you must
1265 be careful to avoid using it as replacement for the C<||> operator.
1266 It usually works out better for flow control than in assignments:
1268 $x = $y or $z; # bug: this is wrong
1269 ($x = $y) or $z; # really means this
1270 $x = $y || $z; # better written this way
1272 However, when it's a list-context assignment and you're trying to use
1273 C<||> for control flow, you probably need "or" so that the assignment
1274 takes higher precedence.
1276 @info = stat($file) || die; # oops, scalar sense of stat!
1277 @info = stat($file) or die; # better, now @info gets its due
1279 Then again, you could always use parentheses.
1281 Binary C<xor> returns the exclusive-OR of the two surrounding expressions.
1282 It cannot short-circuit (of course).
1284 There is no low precedence operator for defined-OR.
1286 =head2 C Operators Missing From Perl
1287 X<operator, missing from perl> X<&> X<*>
1288 X<typecasting> X<(TYPE)>
1290 Here is what C has that Perl doesn't:
1296 Address-of operator. (But see the "\" operator for taking a reference.)
1300 Dereference-address operator. (Perl's prefix dereferencing
1301 operators are typed: $, @, %, and &.)
1305 Type-casting operator.
1309 =head2 Quote and Quote-like Operators
1310 X<operator, quote> X<operator, quote-like> X<q> X<qq> X<qx> X<qw> X<m>
1311 X<qr> X<s> X<tr> X<'> X<''> X<"> X<""> X<//> X<`> X<``> X<<< << >>>
1312 X<escape sequence> X<escape>
1314 While we usually think of quotes as literal values, in Perl they
1315 function as operators, providing various kinds of interpolating and
1316 pattern matching capabilities. Perl provides customary quote characters
1317 for these behaviors, but also provides a way for you to choose your
1318 quote character for any of them. In the following table, a C<{}> represents
1319 any pair of delimiters you choose.
1321 Customary Generic Meaning Interpolates
1324 `` qx{} Command yes*
1326 // m{} Pattern match yes*
1328 s{}{} Substitution yes*
1329 tr{}{} Transliteration no (but see below)
1330 y{}{} Transliteration no (but see below)
1333 * unless the delimiter is ''.
1335 Non-bracketing delimiters use the same character fore and aft, but the four
1336 sorts of ASCII brackets (round, angle, square, curly) all nest, which means
1345 Note, however, that this does not always work for quoting Perl code:
1347 $s = q{ if($x eq "}") ... }; # WRONG
1349 is a syntax error. The C<Text::Balanced> module (standard as of v5.8,
1350 and from CPAN before then) is able to do this properly.
1352 There can be whitespace between the operator and the quoting
1353 characters, except when C<#> is being used as the quoting character.
1354 C<q#foo#> is parsed as the string C<foo>, while C<q #foo#> is the
1355 operator C<q> followed by a comment. Its argument will be taken
1356 from the next line. This allows you to write:
1358 s {foo} # Replace foo
1361 The following escape sequences are available in constructs that interpolate,
1362 and in transliterations:
1363 X<\t> X<\n> X<\r> X<\f> X<\b> X<\a> X<\e> X<\x> X<\0> X<\c> X<\N> X<\N{}>
1366 Sequence Note Description
1372 \a alarm (bell) (BEL)
1374 \x{263A} [1,8] hex char (example: SMILEY)
1375 \x1b [2,8] restricted range hex char (example: ESC)
1376 \N{name} [3] named Unicode character or character sequence
1377 \N{U+263D} [4,8] Unicode character (example: FIRST QUARTER MOON)
1378 \c[ [5] control char (example: chr(27))
1379 \o{23072} [6,8] octal char (example: SMILEY)
1380 \033 [7,8] restricted range octal char (example: ESC)
1386 The result is the character specified by the hexadecimal number between
1387 the braces. See L</[8]> below for details on which character.
1389 Only hexadecimal digits are valid between the braces. If an invalid
1390 character is encountered, a warning will be issued and the invalid
1391 character and all subsequent characters (valid or invalid) within the
1392 braces will be discarded.
1394 If there are no valid digits between the braces, the generated character is
1395 the NULL character (C<\x{00}>). However, an explicit empty brace (C<\x{}>)
1396 will not cause a warning (currently).
1400 The result is the character specified by the hexadecimal number in the range
1401 0x00 to 0xFF. See L</[8]> below for details on which character.
1403 Only hexadecimal digits are valid following C<\x>. When C<\x> is followed
1404 by fewer than two valid digits, any valid digits will be zero-padded. This
1405 means that C<\x7> will be interpreted as C<\x07>, and a lone <\x> will be
1406 interpreted as C<\x00>. Except at the end of a string, having fewer than
1407 two valid digits will result in a warning. Note that although the warning
1408 says the illegal character is ignored, it is only ignored as part of the
1409 escape and will still be used as the subsequent character in the string.
1412 Original Result Warns?
1420 The result is the Unicode character or character sequence given by I<name>.
1425 C<\N{U+I<hexadecimal number>}> means the Unicode character whose Unicode code
1426 point is I<hexadecimal number>.
1430 The character following C<\c> is mapped to some other character as shown in the
1446 \c? chr(127) # (on ASCII platforms)
1448 In other words, it's the character whose code point has had 64 xor'd with
1449 its uppercase. C<\c?> is DELETE on ASCII platforms because
1450 S<C<ord("?") ^ 64>> is 127, and
1451 C<\c@> is NULL because the ord of "@" is 64, so xor'ing 64 itself produces 0.
1453 Also, C<\c\I<X>> yields C< chr(28) . "I<X>"> for any I<X>, but cannot come at the
1454 end of a string, because the backslash would be parsed as escaping the end
1457 On ASCII platforms, the resulting characters from the list above are the
1458 complete set of ASCII controls. This isn't the case on EBCDIC platforms; see
1459 L<perlebcdic/OPERATOR DIFFERENCES> for a full discussion of the
1460 differences between these for ASCII versus EBCDIC platforms.
1462 Use of any other character following the C<"c"> besides those listed above is
1463 discouraged, and as of Perl v5.20, the only characters actually allowed
1464 are the printable ASCII ones, minus the left brace C<"{">. What happens
1465 for any of the allowed other characters is that the value is derived by
1466 xor'ing with the seventh bit, which is 64, and a warning raised if
1467 enabled. Using the non-allowed characters generates a fatal error.
1469 To get platform independent controls, you can use C<\N{...}>.
1473 The result is the character specified by the octal number between the braces.
1474 See L</[8]> below for details on which character.
1476 If a character that isn't an octal digit is encountered, a warning is raised,
1477 and the value is based on the octal digits before it, discarding it and all
1478 following characters up to the closing brace. It is a fatal error if there are
1479 no octal digits at all.
1483 The result is the character specified by the three-digit octal number in the
1484 range 000 to 777 (but best to not use above 077, see next paragraph). See
1485 L</[8]> below for details on which character.
1487 Some contexts allow 2 or even 1 digit, but any usage without exactly
1488 three digits, the first being a zero, may give unintended results. (For
1489 example, in a regular expression it may be confused with a backreference;
1490 see L<perlrebackslash/Octal escapes>.) Starting in Perl 5.14, you may
1491 use C<\o{}> instead, which avoids all these problems. Otherwise, it is best to
1492 use this construct only for ordinals C<\077> and below, remembering to pad to
1493 the left with zeros to make three digits. For larger ordinals, either use
1494 C<\o{}>, or convert to something else, such as to hex and use C<\x{}>
1499 Several constructs above specify a character by a number. That number
1500 gives the character's position in the character set encoding (indexed from 0).
1501 This is called synonymously its ordinal, code position, or code point. Perl
1502 works on platforms that have a native encoding currently of either ASCII/Latin1
1503 or EBCDIC, each of which allow specification of 256 characters. In general, if
1504 the number is 255 (0xFF, 0377) or below, Perl interprets this in the platform's
1505 native encoding. If the number is 256 (0x100, 0400) or above, Perl interprets
1506 it as a Unicode code point and the result is the corresponding Unicode
1507 character. For example C<\x{50}> and C<\o{120}> both are the number 80 in
1508 decimal, which is less than 256, so the number is interpreted in the native
1509 character set encoding. In ASCII the character in the 80th position (indexed
1510 from 0) is the letter "P", and in EBCDIC it is the ampersand symbol "&".
1511 C<\x{100}> and C<\o{400}> are both 256 in decimal, so the number is interpreted
1512 as a Unicode code point no matter what the native encoding is. The name of the
1513 character in the 256th position (indexed by 0) in Unicode is
1514 C<LATIN CAPITAL LETTER A WITH MACRON>.
1516 There are a couple of exceptions to the above rule. S<C<\N{U+I<hex number>}>> is
1517 always interpreted as a Unicode code point, so that C<\N{U+0050}> is "P" even
1518 on EBCDIC platforms. And if L<C<S<use encoding>>|encoding> is in effect, the
1519 number is considered to be in that encoding, and is translated from that into
1520 the platform's native encoding if there is a corresponding native character;
1521 otherwise to Unicode.
1525 B<NOTE>: Unlike C and other languages, Perl has no C<\v> escape sequence for
1526 the vertical tab (VT, which is 11 in both ASCII and EBCDIC), but you may
1529 does have meaning in regular expression patterns in Perl, see L<perlre>.)
1531 The following escape sequences are available in constructs that interpolate,
1532 but not in transliterations.
1533 X<\l> X<\u> X<\L> X<\U> X<\E> X<\Q> X<\F>
1535 \l lowercase next character only
1536 \u titlecase (not uppercase!) next character only
1537 \L lowercase all characters till \E or end of string
1538 \U uppercase all characters till \E or end of string
1539 \F foldcase all characters till \E or end of string
1540 \Q quote (disable) pattern metacharacters till \E or
1542 \E end either case modification or quoted section
1543 (whichever was last seen)
1545 See L<perlfunc/quotemeta> for the exact definition of characters that
1546 are quoted by C<\Q>.
1548 C<\L>, C<\U>, C<\F>, and C<\Q> can stack, in which case you need one
1549 C<\E> for each. For example:
1551 say"This \Qquoting \ubusiness \Uhere isn't quite\E done yet,\E is it?";
1552 This quoting\ Business\ HERE\ ISN\'T\ QUITE\ done\ yet\, is it?
1554 If C<use locale> is in effect (but not C<use locale ':not_characters'>),
1555 the case map used by C<\l>, C<\L>,
1556 C<\u>, and C<\U> is taken from the current locale. See L<perllocale>.
1557 If Unicode (for example, C<\N{}> or code points of 0x100 or
1558 beyond) is being used, the case map used by C<\l>, C<\L>, C<\u>, and
1559 C<\U> is as defined by Unicode. That means that case-mapping
1560 a single character can sometimes produce several characters.
1561 Under C<use locale>, C<\F> produces the same results as C<\L>
1562 for all locales but a UTF-8 one, where it instead uses the Unicode
1565 All systems use the virtual C<"\n"> to represent a line terminator,
1566 called a "newline". There is no such thing as an unvarying, physical
1567 newline character. It is only an illusion that the operating system,
1568 device drivers, C libraries, and Perl all conspire to preserve. Not all
1569 systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
1570 on the ancient Macs (pre-MacOS X) of yesteryear, these used to be reversed,
1571 and on systems without line terminator,
1572 printing C<"\n"> might emit no actual data. In general, use C<"\n"> when
1573 you mean a "newline" for your system, but use the literal ASCII when you
1574 need an exact character. For example, most networking protocols expect
1575 and prefer a CR+LF (C<"\015\012"> or C<"\cM\cJ">) for line terminators,
1576 and although they often accept just C<"\012">, they seldom tolerate just
1577 C<"\015">. If you get in the habit of using C<"\n"> for networking,
1578 you may be burned some day.
1579 X<newline> X<line terminator> X<eol> X<end of line>
1582 For constructs that do interpolate, variables beginning with "C<$>"
1583 or "C<@>" are interpolated. Subscripted variables such as C<$a[3]> or
1584 C<< $href->{key}[0] >> are also interpolated, as are array and hash slices.
1585 But method calls such as C<< $obj->meth >> are not.
1587 Interpolating an array or slice interpolates the elements in order,
1588 separated by the value of C<$">, so is equivalent to interpolating
1589 C<join $", @array>. "Punctuation" arrays such as C<@*> are usually
1590 interpolated only if the name is enclosed in braces C<@{*}>, but the
1591 arrays C<@_>, C<@+>, and C<@-> are interpolated even without braces.
1593 For double-quoted strings, the quoting from C<\Q> is applied after
1594 interpolation and escapes are processed.
1596 "abc\Qfoo\tbar$s\Exyz"
1600 "abc" . quotemeta("foo\tbar$s") . "xyz"
1602 For the pattern of regex operators (C<qr//>, C<m//> and C<s///>),
1603 the quoting from C<\Q> is applied after interpolation is processed,
1604 but before escapes are processed. This allows the pattern to match
1605 literally (except for C<$> and C<@>). For example, the following matches:
1609 Because C<$> or C<@> trigger interpolation, you'll need to use something
1610 like C</\Quser\E\@\Qhost/> to match them literally.
1612 Patterns are subject to an additional level of interpretation as a
1613 regular expression. This is done as a second pass, after variables are
1614 interpolated, so that regular expressions may be incorporated into the
1615 pattern from the variables. If this is not what you want, use C<\Q> to
1616 interpolate a variable literally.
1618 Apart from the behavior described above, Perl does not expand
1619 multiple levels of interpolation. In particular, contrary to the
1620 expectations of shell programmers, back-quotes do I<NOT> interpolate
1621 within double quotes, nor do single quotes impede evaluation of
1622 variables when used within double quotes.
1624 =head2 Regexp Quote-Like Operators
1627 Here are the quote-like operators that apply to pattern
1628 matching and related activities.
1632 =item qr/STRING/msixpodual
1633 X<qr> X</i> X</m> X</o> X</s> X</x> X</p>
1635 This operator quotes (and possibly compiles) its I<STRING> as a regular
1636 expression. I<STRING> is interpolated the same way as I<PATTERN>
1637 in C<m/PATTERN/>. If "'" is used as the delimiter, no interpolation
1638 is done. Returns a Perl value which may be used instead of the
1639 corresponding C</STRING/msixpodual> expression. The returned value is a
1640 normalized version of the original pattern. It magically differs from
1641 a string containing the same characters: C<ref(qr/x/)> returns "Regexp";
1642 however, dereferencing it is not well defined (you currently get the
1643 normalized version of the original pattern, but this may change).
1648 $rex = qr/my.STRING/is;
1649 print $rex; # prints (?si-xm:my.STRING)
1656 The result may be used as a subpattern in a match:
1659 $string =~ /foo${re}bar/; # can be interpolated in other
1661 $string =~ $re; # or used standalone
1662 $string =~ /$re/; # or this way
1664 Since Perl may compile the pattern at the moment of execution of the qr()
1665 operator, using qr() may have speed advantages in some situations,
1666 notably if the result of qr() is used standalone:
1669 my $patterns = shift;
1670 my @compiled = map qr/$_/i, @$patterns;
1673 foreach my $pat (@compiled) {
1674 $success = 1, last if /$pat/;
1680 Precompilation of the pattern into an internal representation at
1681 the moment of qr() avoids a need to recompile the pattern every
1682 time a match C</$pat/> is attempted. (Perl has many other internal
1683 optimizations, but none would be triggered in the above example if
1684 we did not use qr() operator.)
1686 Options (specified by the following modifiers) are:
1688 m Treat string as multiple lines.
1689 s Treat string as single line. (Make . match a newline)
1690 i Do case-insensitive pattern matching.
1691 x Use extended regular expressions.
1692 p When matching preserve a copy of the matched string so
1693 that ${^PREMATCH}, ${^MATCH}, ${^POSTMATCH} will be
1695 o Compile pattern only once.
1696 a ASCII-restrict: Use ASCII for \d, \s, \w; specifying two
1697 a's further restricts /i matching so that no ASCII
1698 character will match a non-ASCII one.
1700 u Use Unicode rules.
1701 d Use Unicode or native charset, as in 5.12 and earlier.
1703 If a precompiled pattern is embedded in a larger pattern then the effect
1704 of "msixpluad" will be propagated appropriately. The effect the "o"
1705 modifier has is not propagated, being restricted to those patterns
1706 explicitly using it.
1708 The last four modifiers listed above, added in Perl 5.14,
1709 control the character set rules, but C</a> is the only one you are likely
1710 to want to specify explicitly; the other three are selected
1711 automatically by various pragmas.
1713 See L<perlre> for additional information on valid syntax for STRING, and
1714 for a detailed look at the semantics of regular expressions. In
1715 particular, all modifiers except the largely obsolete C</o> are further
1716 explained in L<perlre/Modifiers>. C</o> is described in the next section.
1718 =item m/PATTERN/msixpodualgc
1719 X<m> X<operator, match>
1720 X<regexp, options> X<regexp> X<regex, options> X<regex>
1721 X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c>
1723 =item /PATTERN/msixpodualgc
1725 Searches a string for a pattern match, and in scalar context returns
1726 true if it succeeds, false if it fails. If no string is specified
1727 via the C<=~> or C<!~> operator, the $_ string is searched. (The
1728 string specified with C<=~> need not be an lvalue--it may be the
1729 result of an expression evaluation, but remember the C<=~> binds
1730 rather tightly.) See also L<perlre>.
1732 Options are as described in C<qr//> above; in addition, the following match
1733 process modifiers are available:
1735 g Match globally, i.e., find all occurrences.
1736 c Do not reset search position on a failed match when /g is
1739 If "/" is the delimiter then the initial C<m> is optional. With the C<m>
1740 you can use any pair of non-whitespace (ASCII) characters
1741 as delimiters. This is particularly useful for matching path names
1742 that contain "/", to avoid LTS (leaning toothpick syndrome). If "?" is
1743 the delimiter, then a match-only-once rule applies,
1744 described in C<m?PATTERN?> below. If "'" (single quote) is the delimiter,
1745 no interpolation is performed on the PATTERN.
1746 When using a character valid in an identifier, whitespace is required
1749 PATTERN may contain variables, which will be interpolated
1750 every time the pattern search is evaluated, except
1751 for when the delimiter is a single quote. (Note that C<$(>, C<$)>, and
1752 C<$|> are not interpolated because they look like end-of-string tests.)
1753 Perl will not recompile the pattern unless an interpolated
1754 variable that it contains changes. You can force Perl to skip the
1755 test and never recompile by adding a C</o> (which stands for "once")
1756 after the trailing delimiter.
1757 Once upon a time, Perl would recompile regular expressions
1758 unnecessarily, and this modifier was useful to tell it not to do so, in the
1759 interests of speed. But now, the only reasons to use C</o> are one of:
1765 The variables are thousands of characters long and you know that they
1766 don't change, and you need to wring out the last little bit of speed by
1767 having Perl skip testing for that. (There is a maintenance penalty for
1768 doing this, as mentioning C</o> constitutes a promise that you won't
1769 change the variables in the pattern. If you do change them, Perl won't
1774 you want the pattern to use the initial values of the variables
1775 regardless of whether they change or not. (But there are saner ways
1776 of accomplishing this than using C</o>.)
1780 If the pattern contains embedded code, such as
1783 $code = 'foo(?{ $x })';
1786 then perl will recompile each time, even though the pattern string hasn't
1787 changed, to ensure that the current value of C<$x> is seen each time.
1788 Use C</o> if you want to avoid this.
1792 The bottom line is that using C</o> is almost never a good idea.
1794 =item The empty pattern //
1796 If the PATTERN evaluates to the empty string, the last
1797 I<successfully> matched regular expression is used instead. In this
1798 case, only the C<g> and C<c> flags on the empty pattern are honored;
1799 the other flags are taken from the original pattern. If no match has
1800 previously succeeded, this will (silently) act instead as a genuine
1801 empty pattern (which will always match).
1803 Note that it's possible to confuse Perl into thinking C<//> (the empty
1804 regex) is really C<//> (the defined-or operator). Perl is usually pretty
1805 good about this, but some pathological cases might trigger this, such as
1806 C<$x///> (is that C<($x) / (//)> or C<$x // />?) and C<print $fh //>
1807 (C<print $fh(//> or C<print($fh //>?). In all of these examples, Perl
1808 will assume you meant defined-or. If you meant the empty regex, just
1809 use parentheses or spaces to disambiguate, or even prefix the empty
1810 regex with an C<m> (so C<//> becomes C<m//>).
1812 =item Matching in list context
1814 If the C</g> option is not used, C<m//> in list context returns a
1815 list consisting of the subexpressions matched by the parentheses in the
1816 pattern, that is, (C<$1>, C<$2>, C<$3>...) (Note that here C<$1> etc. are
1817 also set). When there are no parentheses in the pattern, the return
1818 value is the list C<(1)> for success.
1819 With or without parentheses, an empty list is returned upon failure.
1823 open(TTY, "+</dev/tty")
1824 || die "can't access /dev/tty: $!";
1826 <TTY> =~ /^y/i && foo(); # do foo if desired
1828 if (/Version: *([0-9.]*)/) { $version = $1; }
1830 next if m#^/usr/spool/uucp#;
1835 print if /$arg/o; # compile only once (no longer needed!)
1838 if (($F1, $F2, $Etc) = ($foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
1840 This last example splits $foo into the first two words and the
1841 remainder of the line, and assigns those three fields to $F1, $F2, and
1842 $Etc. The conditional is true if any variables were assigned; that is,
1843 if the pattern matched.
1845 The C</g> modifier specifies global pattern matching--that is,
1846 matching as many times as possible within the string. How it behaves
1847 depends on the context. In list context, it returns a list of the
1848 substrings matched by any capturing parentheses in the regular
1849 expression. If there are no parentheses, it returns a list of all
1850 the matched strings, as if there were parentheses around the whole
1853 In scalar context, each execution of C<m//g> finds the next match,
1854 returning true if it matches, and false if there is no further match.
1855 The position after the last match can be read or set using the C<pos()>
1856 function; see L<perlfunc/pos>. A failed match normally resets the
1857 search position to the beginning of the string, but you can avoid that
1858 by adding the C</c> modifier (for example, C<m//gc>). Modifying the target
1859 string also resets the search position.
1863 You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
1864 zero-width assertion that matches the exact position where the
1865 previous C<m//g>, if any, left off. Without the C</g> modifier, the
1866 C<\G> assertion still anchors at C<pos()> as it was at the start of
1867 the operation (see L<perlfunc/pos>), but the match is of course only
1868 attempted once. Using C<\G> without C</g> on a target string that has
1869 not previously had a C</g> match applied to it is the same as using
1870 the C<\A> assertion to match the beginning of the string. Note also
1871 that, currently, C<\G> is only properly supported when anchored at the
1872 very beginning of the pattern.
1877 ($one,$five,$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
1881 while ($paragraph = <>) {
1882 while ($paragraph =~ /\p{Ll}['")]*[.!?]+['")]*\s/g) {
1888 Here's another way to check for sentences in a paragraph:
1890 my $sentence_rx = qr{
1891 (?: (?<= ^ ) | (?<= \s ) ) # after start-of-string or
1893 \p{Lu} # capital letter
1894 .*? # a bunch of anything
1895 (?<= \S ) # that ends in non-
1897 (?<! \b [DMS]r ) # but isn't a common abbr.
1901 [.?!] # followed by a sentence
1903 (?= $ | \s ) # in front of end-of-string
1907 while (my $paragraph = <>) {
1908 say "NEW PARAGRAPH";
1910 while ($paragraph =~ /($sentence_rx)/g) {
1911 printf "\tgot sentence %d: <%s>\n", ++$count, $1;
1915 Here's how to use C<m//gc> with C<\G>:
1920 print $1 while /(o)/gc; print "', pos=", pos, "\n";
1922 print $1 if /\G(q)/gc; print "', pos=", pos, "\n";
1924 print $1 while /(p)/gc; print "', pos=", pos, "\n";
1926 print "Final: '$1', pos=",pos,"\n" if /\G(.)/;
1928 The last example should print:
1938 Notice that the final match matched C<q> instead of C<p>, which a match
1939 without the C<\G> anchor would have done. Also note that the final match
1940 did not update C<pos>. C<pos> is only updated on a C</g> match. If the
1941 final match did indeed match C<p>, it's a good bet that you're running a
1942 very old (pre-5.6.0) version of Perl.
1944 A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
1945 combine several regexps like this to process a string part-by-part,
1946 doing different actions depending on which regexp matched. Each
1947 regexp tries to match where the previous one leaves off.
1950 $url = URI::URL->new( "http://example.com/" );
1951 die if $url eq "xXx";
1955 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
1956 print(" lowercase"), redo LOOP
1957 if /\G\p{Ll}+\b[,.;]?\s*/gc;
1958 print(" UPPERCASE"), redo LOOP
1959 if /\G\p{Lu}+\b[,.;]?\s*/gc;
1960 print(" Capitalized"), redo LOOP
1961 if /\G\p{Lu}\p{Ll}+\b[,.;]?\s*/gc;
1962 print(" MiXeD"), redo LOOP if /\G\pL+\b[,.;]?\s*/gc;
1963 print(" alphanumeric"), redo LOOP
1964 if /\G[\p{Alpha}\pN]+\b[,.;]?\s*/gc;
1965 print(" line-noise"), redo LOOP if /\G\W+/gc;
1966 print ". That's all!\n";
1969 Here is the output (split into several lines):
1971 line-noise lowercase line-noise UPPERCASE line-noise UPPERCASE
1972 line-noise lowercase line-noise lowercase line-noise lowercase
1973 lowercase line-noise lowercase lowercase line-noise lowercase
1974 lowercase line-noise MiXeD line-noise. That's all!
1976 =item m?PATTERN?msixpodualgc
1977 X<?> X<operator, match-once>
1979 =item ?PATTERN?msixpodualgc
1981 This is just like the C<m/PATTERN/> search, except that it matches
1982 only once between calls to the reset() operator. This is a useful
1983 optimization when you want to see only the first occurrence of
1984 something in each file of a set of files, for instance. Only C<m??>
1985 patterns local to the current package are reset.
1989 # blank line between header and body
1992 reset if eof; # clear m?? status for next file
1995 Another example switched the first "latin1" encoding it finds
1996 to "utf8" in a pod file:
1998 s//utf8/ if m? ^ =encoding \h+ \K latin1 ?x;
2000 The match-once behavior is controlled by the match delimiter being
2001 C<?>; with any other delimiter this is the normal C<m//> operator.
2003 For historical reasons, the leading C<m> in C<m?PATTERN?> is optional,
2004 but the resulting C<?PATTERN?> syntax is deprecated, will warn on
2005 usage and might be removed from a future stable release of Perl (without
2008 =item s/PATTERN/REPLACEMENT/msixpodualgcer
2009 X<substitute> X<substitution> X<replace> X<regexp, replace>
2010 X<regexp, substitute> X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c> X</e> X</r>
2012 Searches a string for a pattern, and if found, replaces that pattern
2013 with the replacement text and returns the number of substitutions
2014 made. Otherwise it returns false (specifically, the empty string).
2016 If the C</r> (non-destructive) option is used then it runs the
2017 substitution on a copy of the string and instead of returning the
2018 number of substitutions, it returns the copy whether or not a
2019 substitution occurred. The original string is never changed when
2020 C</r> is used. The copy will always be a plain string, even if the
2021 input is an object or a tied variable.
2023 If no string is specified via the C<=~> or C<!~> operator, the C<$_>
2024 variable is searched and modified. Unless the C</r> option is used,
2025 the string specified must be a scalar variable, an array element, a
2026 hash element, or an assignment to one of those; that is, some sort of
2029 If the delimiter chosen is a single quote, no interpolation is
2030 done on either the PATTERN or the REPLACEMENT. Otherwise, if the
2031 PATTERN contains a $ that looks like a variable rather than an
2032 end-of-string test, the variable will be interpolated into the pattern
2033 at run-time. If you want the pattern compiled only once the first time
2034 the variable is interpolated, use the C</o> option. If the pattern
2035 evaluates to the empty string, the last successfully executed regular
2036 expression is used instead. See L<perlre> for further explanation on these.
2038 Options are as with m// with the addition of the following replacement
2041 e Evaluate the right side as an expression.
2042 ee Evaluate the right side as a string then eval the
2044 r Return substitution and leave the original string
2047 Any non-whitespace delimiter may replace the slashes. Add space after
2048 the C<s> when using a character allowed in identifiers. If single quotes
2049 are used, no interpretation is done on the replacement string (the C</e>
2050 modifier overrides this, however). Note that Perl treats backticks
2051 as normal delimiters; the replacement text is not evaluated as a command.
2052 If the PATTERN is delimited by bracketing quotes, the REPLACEMENT has
2053 its own pair of quotes, which may or may not be bracketing quotes, for example,
2054 C<s(foo)(bar)> or C<< s<foo>/bar/ >>. A C</e> will cause the
2055 replacement portion to be treated as a full-fledged Perl expression
2056 and evaluated right then and there. It is, however, syntax checked at
2057 compile-time. A second C<e> modifier will cause the replacement portion
2058 to be C<eval>ed before being run as a Perl expression.
2062 s/\bgreen\b/mauve/g; # don't change wintergreen
2064 $path =~ s|/usr/bin|/usr/local/bin|;
2066 s/Login: $foo/Login: $bar/; # run-time pattern
2068 ($foo = $bar) =~ s/this/that/; # copy first, then
2070 ($foo = "$bar") =~ s/this/that/; # convert to string,
2072 $foo = $bar =~ s/this/that/r; # Same as above using /r
2073 $foo = $bar =~ s/this/that/r
2074 =~ s/that/the other/r; # Chained substitutes
2076 @foo = map { s/this/that/r } @bar # /r is very useful in
2079 $count = ($paragraph =~ s/Mister\b/Mr./g); # get change-cnt
2082 s/\d+/$&*2/e; # yields 'abc246xyz'
2083 s/\d+/sprintf("%5d",$&)/e; # yields 'abc 246xyz'
2084 s/\w/$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
2086 s/%(.)/$percent{$1}/g; # change percent escapes; no /e
2087 s/%(.)/$percent{$1} || $&/ge; # expr now, so /e
2088 s/^=(\w+)/pod($1)/ge; # use function call
2091 $x = s/abc/def/r; # $x is 'def123xyz' and
2092 # $_ remains 'abc123xyz'.
2094 # expand variables in $_, but dynamics only, using
2095 # symbolic dereferencing
2098 # Add one to the value of any numbers in the string
2101 # Titlecase words in the last 30 characters only
2102 substr($str, -30) =~ s/\b(\p{Alpha}+)\b/\u\L$1/g;
2104 # This will expand any embedded scalar variable
2105 # (including lexicals) in $_ : First $1 is interpolated
2106 # to the variable name, and then evaluated
2109 # Delete (most) C comments.
2111 /\* # Match the opening delimiter.
2112 .*? # Match a minimal number of characters.
2113 \*/ # Match the closing delimiter.
2116 s/^\s*(.*?)\s*$/$1/; # trim whitespace in $_,
2119 for ($variable) { # trim whitespace in $variable,
2125 s/([^ ]*) *([^ ]*)/$2 $1/; # reverse 1st two fields
2127 Note the use of $ instead of \ in the last example. Unlike
2128 B<sed>, we use the \<I<digit>> form in only the left hand side.
2129 Anywhere else it's $<I<digit>>.
2131 Occasionally, you can't use just a C</g> to get all the changes
2132 to occur that you might want. Here are two common cases:
2134 # put commas in the right places in an integer
2135 1 while s/(\d)(\d\d\d)(?!\d)/$1,$2/g;
2137 # expand tabs to 8-column spacing
2138 1 while s/\t+/' ' x (length($&)*8 - length($`)%8)/e;
2142 =head2 Quote-Like Operators
2143 X<operator, quote-like>
2148 X<q> X<quote, single> X<'> X<''>
2152 A single-quoted, literal string. A backslash represents a backslash
2153 unless followed by the delimiter or another backslash, in which case
2154 the delimiter or backslash is interpolated.
2156 $foo = q!I said, "You said, 'She said it.'"!;
2157 $bar = q('This is it.');
2158 $baz = '\n'; # a two-character string
2161 X<qq> X<quote, double> X<"> X<"">
2165 A double-quoted, interpolated string.
2168 (*** The previous line contains the naughty word "$1".\n)
2169 if /\b(tcl|java|python)\b/i; # :-)
2170 $baz = "\n"; # a one-character string
2173 X<qx> X<`> X<``> X<backtick>
2177 A string which is (possibly) interpolated and then executed as a
2178 system command with F</bin/sh> or its equivalent. Shell wildcards,
2179 pipes, and redirections will be honored. The collected standard
2180 output of the command is returned; standard error is unaffected. In
2181 scalar context, it comes back as a single (potentially multi-line)
2182 string, or undef if the command failed. In list context, returns a
2183 list of lines (however you've defined lines with $/ or
2184 $INPUT_RECORD_SEPARATOR), or an empty list if the command failed.
2186 Because backticks do not affect standard error, use shell file descriptor
2187 syntax (assuming the shell supports this) if you care to address this.
2188 To capture a command's STDERR and STDOUT together:
2190 $output = `cmd 2>&1`;
2192 To capture a command's STDOUT but discard its STDERR:
2194 $output = `cmd 2>/dev/null`;
2196 To capture a command's STDERR but discard its STDOUT (ordering is
2199 $output = `cmd 2>&1 1>/dev/null`;
2201 To exchange a command's STDOUT and STDERR in order to capture the STDERR
2202 but leave its STDOUT to come out the old STDERR:
2204 $output = `cmd 3>&1 1>&2 2>&3 3>&-`;
2206 To read both a command's STDOUT and its STDERR separately, it's easiest
2207 to redirect them separately to files, and then read from those files
2208 when the program is done:
2210 system("program args 1>program.stdout 2>program.stderr");
2212 The STDIN filehandle used by the command is inherited from Perl's STDIN.
2215 open(SPLAT, "stuff") || die "can't open stuff: $!";
2216 open(STDIN, "<&SPLAT") || die "can't dupe SPLAT: $!";
2217 print STDOUT `sort`;
2219 will print the sorted contents of the file named F<"stuff">.
2221 Using single-quote as a delimiter protects the command from Perl's
2222 double-quote interpolation, passing it on to the shell instead:
2224 $perl_info = qx(ps $$); # that's Perl's $$
2225 $shell_info = qx'ps $$'; # that's the new shell's $$
2227 How that string gets evaluated is entirely subject to the command
2228 interpreter on your system. On most platforms, you will have to protect
2229 shell metacharacters if you want them treated literally. This is in
2230 practice difficult to do, as it's unclear how to escape which characters.
2231 See L<perlsec> for a clean and safe example of a manual fork() and exec()
2232 to emulate backticks safely.
2234 On some platforms (notably DOS-like ones), the shell may not be
2235 capable of dealing with multiline commands, so putting newlines in
2236 the string may not get you what you want. You may be able to evaluate
2237 multiple commands in a single line by separating them with the command
2238 separator character, if your shell supports that (for example, C<;> on
2239 many Unix shells and C<&> on the Windows NT C<cmd> shell).
2241 Perl will attempt to flush all files opened for
2242 output before starting the child process, but this may not be supported
2243 on some platforms (see L<perlport>). To be safe, you may need to set
2244 C<$|> ($AUTOFLUSH in English) or call the C<autoflush()> method of
2245 C<IO::Handle> on any open handles.
2247 Beware that some command shells may place restrictions on the length
2248 of the command line. You must ensure your strings don't exceed this
2249 limit after any necessary interpolations. See the platform-specific
2250 release notes for more details about your particular environment.
2252 Using this operator can lead to programs that are difficult to port,
2253 because the shell commands called vary between systems, and may in
2254 fact not be present at all. As one example, the C<type> command under
2255 the POSIX shell is very different from the C<type> command under DOS.
2256 That doesn't mean you should go out of your way to avoid backticks
2257 when they're the right way to get something done. Perl was made to be
2258 a glue language, and one of the things it glues together is commands.
2259 Just understand what you're getting yourself into.
2261 See L</"I/O Operators"> for more discussion.
2264 X<qw> X<quote, list> X<quote, words>
2266 Evaluates to a list of the words extracted out of STRING, using embedded
2267 whitespace as the word delimiters. It can be understood as being roughly
2270 split(" ", q/STRING/);
2272 the differences being that it generates a real list at compile time, and
2273 in scalar context it returns the last element in the list. So
2278 is semantically equivalent to the list:
2282 Some frequently seen examples:
2284 use POSIX qw( setlocale localeconv )
2285 @EXPORT = qw( foo bar baz );
2287 A common mistake is to try to separate the words with comma or to
2288 put comments into a multi-line C<qw>-string. For this reason, the
2289 C<use warnings> pragma and the B<-w> switch (that is, the C<$^W> variable)
2290 produces warnings if the STRING contains the "," or the "#" character.
2292 =item tr/SEARCHLIST/REPLACEMENTLIST/cdsr
2293 X<tr> X<y> X<transliterate> X</c> X</d> X</s>
2295 =item y/SEARCHLIST/REPLACEMENTLIST/cdsr
2297 Transliterates all occurrences of the characters found in the search list
2298 with the corresponding character in the replacement list. It returns
2299 the number of characters replaced or deleted. If no string is
2300 specified via the C<=~> or C<!~> operator, the $_ string is transliterated.
2302 If the C</r> (non-destructive) option is present, a new copy of the string
2303 is made and its characters transliterated, and this copy is returned no
2304 matter whether it was modified or not: the original string is always
2305 left unchanged. The new copy is always a plain string, even if the input
2306 string is an object or a tied variable.
2308 Unless the C</r> option is used, the string specified with C<=~> must be a
2309 scalar variable, an array element, a hash element, or an assignment to one
2310 of those; in other words, an lvalue.
2312 A character range may be specified with a hyphen, so C<tr/A-J/0-9/>
2313 does the same replacement as C<tr/ACEGIBDFHJ/0246813579/>.
2314 For B<sed> devotees, C<y> is provided as a synonym for C<tr>. If the
2315 SEARCHLIST is delimited by bracketing quotes, the REPLACEMENTLIST has
2316 its own pair of quotes, which may or may not be bracketing quotes;
2317 for example, C<tr[aeiouy][yuoiea]> or C<tr(+\-*/)/ABCD/>.
2319 Note that C<tr> does B<not> do regular expression character classes such as
2320 C<\d> or C<\pL>. The C<tr> operator is not equivalent to the tr(1)
2321 utility. If you want to map strings between lower/upper cases, see
2322 L<perlfunc/lc> and L<perlfunc/uc>, and in general consider using the C<s>
2323 operator if you need regular expressions. The C<\U>, C<\u>, C<\L>, and
2324 C<\l> string-interpolation escapes on the right side of a substitution
2325 operator will perform correct case-mappings, but C<tr[a-z][A-Z]> will not
2326 (except sometimes on legacy 7-bit data).
2328 Note also that the whole range idea is rather unportable between
2329 character sets--and even within character sets they may cause results
2330 you probably didn't expect. A sound principle is to use only ranges
2331 that begin from and end at either alphabets of equal case (a-e, A-E),
2332 or digits (0-4). Anything else is unsafe. If in doubt, spell out the
2333 character sets in full.
2337 c Complement the SEARCHLIST.
2338 d Delete found but unreplaced characters.
2339 s Squash duplicate replaced characters.
2340 r Return the modified string and leave the original string
2343 If the C</c> modifier is specified, the SEARCHLIST character set
2344 is complemented. If the C</d> modifier is specified, any characters
2345 specified by SEARCHLIST not found in REPLACEMENTLIST are deleted.
2346 (Note that this is slightly more flexible than the behavior of some
2347 B<tr> programs, which delete anything they find in the SEARCHLIST,
2348 period.) If the C</s> modifier is specified, sequences of characters
2349 that were transliterated to the same character are squashed down
2350 to a single instance of the character.
2352 If the C</d> modifier is used, the REPLACEMENTLIST is always interpreted
2353 exactly as specified. Otherwise, if the REPLACEMENTLIST is shorter
2354 than the SEARCHLIST, the final character is replicated till it is long
2355 enough. If the REPLACEMENTLIST is empty, the SEARCHLIST is replicated.
2356 This latter is useful for counting characters in a class or for
2357 squashing character sequences in a class.
2361 $ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case ASCII
2363 $cnt = tr/*/*/; # count the stars in $_
2365 $cnt = $sky =~ tr/*/*/; # count the stars in $sky
2367 $cnt = tr/0-9//; # count the digits in $_
2369 tr/a-zA-Z//s; # bookkeeper -> bokeper
2371 ($HOST = $host) =~ tr/a-z/A-Z/;
2372 $HOST = $host =~ tr/a-z/A-Z/r; # same thing
2374 $HOST = $host =~ tr/a-z/A-Z/r # chained with s///r
2377 tr/a-zA-Z/ /cs; # change non-alphas to single space
2379 @stripped = map tr/a-zA-Z/ /csr, @original;
2383 [\000-\177]; # wickedly delete 8th bit
2385 If multiple transliterations are given for a character, only the
2390 will transliterate any A to X.
2392 Because the transliteration table is built at compile time, neither
2393 the SEARCHLIST nor the REPLACEMENTLIST are subjected to double quote
2394 interpolation. That means that if you want to use variables, you
2397 eval "tr/$oldlist/$newlist/";
2400 eval "tr/$oldlist/$newlist/, 1" or die $@;
2403 X<here-doc> X<heredoc> X<here-document> X<<< << >>>
2405 A line-oriented form of quoting is based on the shell "here-document"
2406 syntax. Following a C<< << >> you specify a string to terminate
2407 the quoted material, and all lines following the current line down to
2408 the terminating string are the value of the item.
2410 The terminating string may be either an identifier (a word), or some
2411 quoted text. An unquoted identifier works like double quotes.
2412 There may not be a space between the C<< << >> and the identifier,
2413 unless the identifier is explicitly quoted. (If you put a space it
2414 will be treated as a null identifier, which is valid, and matches the
2415 first empty line.) The terminating string must appear by itself
2416 (unquoted and with no surrounding whitespace) on the terminating line.
2418 If the terminating string is quoted, the type of quotes used determine
2419 the treatment of the text.
2425 Double quotes indicate that the text will be interpolated using exactly
2426 the same rules as normal double quoted strings.
2429 The price is $Price.
2432 print << "EOF"; # same as above
2433 The price is $Price.
2439 Single quotes indicate the text is to be treated literally with no
2440 interpolation of its content. This is similar to single quoted
2441 strings except that backslashes have no special meaning, with C<\\>
2442 being treated as two backslashes and not one as they would in every
2443 other quoting construct.
2445 Just as in the shell, a backslashed bareword following the C<<< << >>>
2446 means the same thing as a single-quoted string does:
2448 $cost = <<'VISTA'; # hasta la ...
2449 That'll be $10 please, ma'am.
2452 $cost = <<\VISTA; # Same thing!
2453 That'll be $10 please, ma'am.
2456 This is the only form of quoting in perl where there is no need
2457 to worry about escaping content, something that code generators
2458 can and do make good use of.
2462 The content of the here doc is treated just as it would be if the
2463 string were embedded in backticks. Thus the content is interpolated
2464 as though it were double quoted and then executed via the shell, with
2465 the results of the execution returned.
2467 print << `EOC`; # execute command and get results
2473 It is possible to stack multiple here-docs in a row:
2475 print <<"foo", <<"bar"; # you can stack them
2481 myfunc(<< "THIS", 23, <<'THAT');
2488 Just don't forget that you have to put a semicolon on the end
2489 to finish the statement, as Perl doesn't know you're not going to
2497 If you want to remove the line terminator from your here-docs,
2500 chomp($string = <<'END');
2504 If you want your here-docs to be indented with the rest of the code,
2505 you'll need to remove leading whitespace from each line manually:
2507 ($quote = <<'FINIS') =~ s/^\s+//gm;
2508 The Road goes ever on and on,
2509 down from the door where it began.
2512 If you use a here-doc within a delimited construct, such as in C<s///eg>,
2513 the quoted material must still come on the line following the
2514 C<<< <<FOO >>> marker, which means it may be inside the delimited
2522 It works this way as of Perl 5.18. Historically, it was inconsistent, and
2523 you would have to write
2530 outside of string evals.
2532 Additionally, quoting rules for the end-of-string identifier are
2533 unrelated to Perl's quoting rules. C<q()>, C<qq()>, and the like are not
2534 supported in place of C<''> and C<"">, and the only interpolation is for
2535 backslashing the quoting character:
2537 print << "abc\"def";
2541 Finally, quoted strings cannot span multiple lines. The general rule is
2542 that the identifier must be a string literal. Stick with that, and you
2547 =head2 Gory details of parsing quoted constructs
2548 X<quote, gory details>
2550 When presented with something that might have several different
2551 interpretations, Perl uses the B<DWIM> (that's "Do What I Mean")
2552 principle to pick the most probable interpretation. This strategy
2553 is so successful that Perl programmers often do not suspect the
2554 ambivalence of what they write. But from time to time, Perl's
2555 notions differ substantially from what the author honestly meant.
2557 This section hopes to clarify how Perl handles quoted constructs.
2558 Although the most common reason to learn this is to unravel labyrinthine
2559 regular expressions, because the initial steps of parsing are the
2560 same for all quoting operators, they are all discussed together.
2562 The most important Perl parsing rule is the first one discussed
2563 below: when processing a quoted construct, Perl first finds the end
2564 of that construct, then interprets its contents. If you understand
2565 this rule, you may skip the rest of this section on the first
2566 reading. The other rules are likely to contradict the user's
2567 expectations much less frequently than this first one.
2569 Some passes discussed below are performed concurrently, but because
2570 their results are the same, we consider them individually. For different
2571 quoting constructs, Perl performs different numbers of passes, from
2572 one to four, but these passes are always performed in the same order.
2576 =item Finding the end
2578 The first pass is finding the end of the quoted construct, where
2579 the information about the delimiters is used in parsing.
2580 During this search, text between the starting and ending delimiters
2581 is copied to a safe location. The text copied gets delimiter-independent.
2583 If the construct is a here-doc, the ending delimiter is a line
2584 that has a terminating string as the content. Therefore C<<<EOF> is
2585 terminated by C<EOF> immediately followed by C<"\n"> and starting
2586 from the first column of the terminating line.
2587 When searching for the terminating line of a here-doc, nothing
2588 is skipped. In other words, lines after the here-doc syntax
2589 are compared with the terminating string line by line.
2591 For the constructs except here-docs, single characters are used as starting
2592 and ending delimiters. If the starting delimiter is an opening punctuation
2593 (that is C<(>, C<[>, C<{>, or C<< < >>), the ending delimiter is the
2594 corresponding closing punctuation (that is C<)>, C<]>, C<}>, or C<< > >>).
2595 If the starting delimiter is an unpaired character like C</> or a closing
2596 punctuation, the ending delimiter is same as the starting delimiter.
2597 Therefore a C</> terminates a C<qq//> construct, while a C<]> terminates
2598 both C<qq[]> and C<qq]]> constructs.
2600 When searching for single-character delimiters, escaped delimiters
2601 and C<\\> are skipped. For example, while searching for terminating C</>,
2602 combinations of C<\\> and C<\/> are skipped. If the delimiters are
2603 bracketing, nested pairs are also skipped. For example, while searching
2604 for closing C<]> paired with the opening C<[>, combinations of C<\\>, C<\]>,
2605 and C<\[> are all skipped, and nested C<[> and C<]> are skipped as well.
2606 However, when backslashes are used as the delimiters (like C<qq\\> and
2607 C<tr\\\>), nothing is skipped.
2608 During the search for the end, backslashes that escape delimiters or
2609 other backslashes are removed (exactly speaking, they are not copied to the
2612 For constructs with three-part delimiters (C<s///>, C<y///>, and
2613 C<tr///>), the search is repeated once more.
2614 If the first delimiter is not an opening punctuation, the three delimiters must
2615 be the same, such as C<s!!!> and C<tr)))>,
2616 in which case the second delimiter
2617 terminates the left part and starts the right part at once.
2618 If the left part is delimited by bracketing punctuation (that is C<()>,
2619 C<[]>, C<{}>, or C<< <> >>), the right part needs another pair of
2620 delimiters such as C<s(){}> and C<tr[]//>. In these cases, whitespace
2621 and comments are allowed between the two parts, though the comment must follow
2622 at least one whitespace character; otherwise a character expected as the
2623 start of the comment may be regarded as the starting delimiter of the right part.
2625 During this search no attention is paid to the semantics of the construct.
2628 "$hash{"$foo/$bar"}"
2633 bar # NOT a comment, this slash / terminated m//!
2636 do not form legal quoted expressions. The quoted part ends on the
2637 first C<"> and C</>, and the rest happens to be a syntax error.
2638 Because the slash that terminated C<m//> was followed by a C<SPACE>,
2639 the example above is not C<m//x>, but rather C<m//> with no C</x>
2640 modifier. So the embedded C<#> is interpreted as a literal C<#>.
2642 Also no attention is paid to C<\c\> (multichar control char syntax) during
2643 this search. Thus the second C<\> in C<qq/\c\/> is interpreted as a part
2644 of C<\/>, and the following C</> is not recognized as a delimiter.
2645 Instead, use C<\034> or C<\x1c> at the end of quoted constructs.
2650 The next step is interpolation in the text obtained, which is now
2651 delimiter-independent. There are multiple cases.
2657 No interpolation is performed.
2658 Note that the combination C<\\> is left intact, since escaped delimiters
2659 are not available for here-docs.
2661 =item C<m''>, the pattern of C<s'''>
2663 No interpolation is performed at this stage.
2664 Any backslashed sequences including C<\\> are treated at the stage
2665 to L</"parsing regular expressions">.
2667 =item C<''>, C<q//>, C<tr'''>, C<y'''>, the replacement of C<s'''>
2669 The only interpolation is removal of C<\> from pairs of C<\\>.
2670 Therefore C<-> in C<tr'''> and C<y'''> is treated literally
2671 as a hyphen and no character range is available.
2672 C<\1> in the replacement of C<s'''> does not work as C<$1>.
2674 =item C<tr///>, C<y///>
2676 No variable interpolation occurs. String modifying combinations for
2677 case and quoting such as C<\Q>, C<\U>, and C<\E> are not recognized.
2678 The other escape sequences such as C<\200> and C<\t> and backslashed
2679 characters such as C<\\> and C<\-> are converted to appropriate literals.
2680 The character C<-> is treated specially and therefore C<\-> is treated
2683 =item C<"">, C<``>, C<qq//>, C<qx//>, C<< <file*glob> >>, C<<<"EOF">
2685 C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\F> (possibly paired with C<\E>) are
2686 converted to corresponding Perl constructs. Thus, C<"$foo\Qbaz$bar">
2687 is converted to C<$foo . (quotemeta("baz" . $bar))> internally.
2688 The other escape sequences such as C<\200> and C<\t> and backslashed
2689 characters such as C<\\> and C<\-> are replaced with appropriate
2692 Let it be stressed that I<whatever falls between C<\Q> and C<\E>>
2693 is interpolated in the usual way. Something like C<"\Q\\E"> has
2694 no C<\E> inside. Instead, it has C<\Q>, C<\\>, and C<E>, so the
2695 result is the same as for C<"\\\\E">. As a general rule, backslashes
2696 between C<\Q> and C<\E> may lead to counterintuitive results. So,
2697 C<"\Q\t\E"> is converted to C<quotemeta("\t")>, which is the same
2698 as C<"\\\t"> (since TAB is not alphanumeric). Note also that:
2703 may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
2705 Interpolated scalars and arrays are converted internally to the C<join> and
2706 C<.> catenation operations. Thus, C<"$foo XXX '@arr'"> becomes:
2708 $foo . " XXX '" . (join $", @arr) . "'";
2710 All operations above are performed simultaneously, left to right.
2712 Because the result of C<"\Q STRING \E"> has all metacharacters
2713 quoted, there is no way to insert a literal C<$> or C<@> inside a
2714 C<\Q\E> pair. If protected by C<\>, C<$> will be quoted to became
2715 C<"\\\$">; if not, it is interpreted as the start of an interpolated
2718 Note also that the interpolation code needs to make a decision on
2719 where the interpolated scalar ends. For instance, whether
2720 C<< "a $x -> {c}" >> really means:
2722 "a " . $x . " -> {c}";
2728 Most of the time, the longest possible text that does not include
2729 spaces between components and which contains matching braces or
2730 brackets. because the outcome may be determined by voting based
2731 on heuristic estimators, the result is not strictly predictable.
2732 Fortunately, it's usually correct for ambiguous cases.
2734 =item the replacement of C<s///>
2736 Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\F> and interpolation
2737 happens as with C<qq//> constructs.
2739 It is at this step that C<\1> is begrudgingly converted to C<$1> in
2740 the replacement text of C<s///>, in order to correct the incorrigible
2741 I<sed> hackers who haven't picked up the saner idiom yet. A warning
2742 is emitted if the C<use warnings> pragma or the B<-w> command-line flag
2743 (that is, the C<$^W> variable) was set.
2745 =item C<RE> in C<?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
2747 Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\F>, C<\E>,
2748 and interpolation happens (almost) as with C<qq//> constructs.
2750 Processing of C<\N{...}> is also done here, and compiled into an intermediate
2751 form for the regex compiler. (This is because, as mentioned below, the regex
2752 compilation may be done at execution time, and C<\N{...}> is a compile-time
2755 However any other combinations of C<\> followed by a character
2756 are not substituted but only skipped, in order to parse them
2757 as regular expressions at the following step.
2758 As C<\c> is skipped at this step, C<@> of C<\c@> in RE is possibly
2759 treated as an array symbol (for example C<@foo>),
2760 even though the same text in C<qq//> gives interpolation of C<\c@>.
2762 Code blocks such as C<(?{BLOCK})> are handled by temporarily passing control
2763 back to the perl parser, in a similar way that an interpolated array
2764 subscript expression such as C<"foo$array[1+f("[xyz")]bar"> would be.
2766 Moreover, inside C<(?{BLOCK})>, C<(?# comment )>, and
2767 a C<#>-comment in a C<//x>-regular expression, no processing is
2768 performed whatsoever. This is the first step at which the presence
2769 of the C<//x> modifier is relevant.
2771 Interpolation in patterns has several quirks: C<$|>, C<$(>, C<$)>, C<@+>
2772 and C<@-> are not interpolated, and constructs C<$var[SOMETHING]> are
2773 voted (by several different estimators) to be either an array element
2774 or C<$var> followed by an RE alternative. This is where the notation
2775 C<${arr[$bar]}> comes handy: C</${arr[0-9]}/> is interpreted as
2776 array element C<-9>, not as a regular expression from the variable
2777 C<$arr> followed by a digit, which would be the interpretation of
2778 C</$arr[0-9]/>. Since voting among different estimators may occur,
2779 the result is not predictable.
2781 The lack of processing of C<\\> creates specific restrictions on
2782 the post-processed text. If the delimiter is C</>, one cannot get
2783 the combination C<\/> into the result of this step. C</> will
2784 finish the regular expression, C<\/> will be stripped to C</> on
2785 the previous step, and C<\\/> will be left as is. Because C</> is
2786 equivalent to C<\/> inside a regular expression, this does not
2787 matter unless the delimiter happens to be character special to the
2788 RE engine, such as in C<s*foo*bar*>, C<m[foo]>, or C<?foo?>; or an
2789 alphanumeric char, as in:
2793 In the RE above, which is intentionally obfuscated for illustration, the
2794 delimiter is C<m>, the modifier is C<mx>, and after delimiter-removal the
2795 RE is the same as for C<m/ ^ a \s* b /mx>. There's more than one
2796 reason you're encouraged to restrict your delimiters to non-alphanumeric,
2797 non-whitespace choices.
2801 This step is the last one for all constructs except regular expressions,
2802 which are processed further.
2804 =item parsing regular expressions
2807 Previous steps were performed during the compilation of Perl code,
2808 but this one happens at run time, although it may be optimized to
2809 be calculated at compile time if appropriate. After preprocessing
2810 described above, and possibly after evaluation if concatenation,
2811 joining, casing translation, or metaquoting are involved, the
2812 resulting I<string> is passed to the RE engine for compilation.
2814 Whatever happens in the RE engine might be better discussed in L<perlre>,
2815 but for the sake of continuity, we shall do so here.
2817 This is another step where the presence of the C<//x> modifier is
2818 relevant. The RE engine scans the string from left to right and
2819 converts it to a finite automaton.
2821 Backslashed characters are either replaced with corresponding
2822 literal strings (as with C<\{>), or else they generate special nodes
2823 in the finite automaton (as with C<\b>). Characters special to the
2824 RE engine (such as C<|>) generate corresponding nodes or groups of
2825 nodes. C<(?#...)> comments are ignored. All the rest is either
2826 converted to literal strings to match, or else is ignored (as is
2827 whitespace and C<#>-style comments if C<//x> is present).
2829 Parsing of the bracketed character class construct, C<[...]>, is
2830 rather different than the rule used for the rest of the pattern.
2831 The terminator of this construct is found using the same rules as
2832 for finding the terminator of a C<{}>-delimited construct, the only
2833 exception being that C<]> immediately following C<[> is treated as
2834 though preceded by a backslash.
2836 The terminator of runtime C<(?{...})> is found by temporarily switching
2837 control to the perl parser, which should stop at the point where the
2838 logically balancing terminating C<}> is found.
2840 It is possible to inspect both the string given to RE engine and the
2841 resulting finite automaton. See the arguments C<debug>/C<debugcolor>
2842 in the C<use L<re>> pragma, as well as Perl's B<-Dr> command-line
2843 switch documented in L<perlrun/"Command Switches">.
2845 =item Optimization of regular expressions
2846 X<regexp, optimization>
2848 This step is listed for completeness only. Since it does not change
2849 semantics, details of this step are not documented and are subject
2850 to change without notice. This step is performed over the finite
2851 automaton that was generated during the previous pass.
2853 It is at this stage that C<split()> silently optimizes C</^/> to
2858 =head2 I/O Operators
2859 X<operator, i/o> X<operator, io> X<io> X<while> X<filehandle>
2860 X<< <> >> X<< <<>> >> X<@ARGV>
2862 There are several I/O operators you should know about.
2864 A string enclosed by backticks (grave accents) first undergoes
2865 double-quote interpolation. It is then interpreted as an external
2866 command, and the output of that command is the value of the
2867 backtick string, like in a shell. In scalar context, a single string
2868 consisting of all output is returned. In list context, a list of
2869 values is returned, one per line of output. (You can set C<$/> to use
2870 a different line terminator.) The command is executed each time the
2871 pseudo-literal is evaluated. The status value of the command is
2872 returned in C<$?> (see L<perlvar> for the interpretation of C<$?>).
2873 Unlike in B<csh>, no translation is done on the return data--newlines
2874 remain newlines. Unlike in any of the shells, single quotes do not
2875 hide variable names in the command from interpretation. To pass a
2876 literal dollar-sign through to the shell you need to hide it with a
2877 backslash. The generalized form of backticks is C<qx//>. (Because
2878 backticks always undergo shell expansion as well, see L<perlsec> for
2880 X<qx> X<`> X<``> X<backtick> X<glob>
2882 In scalar context, evaluating a filehandle in angle brackets yields
2883 the next line from that file (the newline, if any, included), or
2884 C<undef> at end-of-file or on error. When C<$/> is set to C<undef>
2885 (sometimes known as file-slurp mode) and the file is empty, it
2886 returns C<''> the first time, followed by C<undef> subsequently.
2888 Ordinarily you must assign the returned value to a variable, but
2889 there is one situation where an automatic assignment happens. If
2890 and only if the input symbol is the only thing inside the conditional
2891 of a C<while> statement (even if disguised as a C<for(;;)> loop),
2892 the value is automatically assigned to the global variable $_,
2893 destroying whatever was there previously. (This may seem like an
2894 odd thing to you, but you'll use the construct in almost every Perl
2895 script you write.) The $_ variable is not implicitly localized.
2896 You'll have to put a C<local $_;> before the loop if you want that
2899 The following lines are equivalent:
2901 while (defined($_ = <STDIN>)) { print; }
2902 while ($_ = <STDIN>) { print; }
2903 while (<STDIN>) { print; }
2904 for (;<STDIN>;) { print; }
2905 print while defined($_ = <STDIN>);
2906 print while ($_ = <STDIN>);
2907 print while <STDIN>;
2909 This also behaves similarly, but assigns to a lexical variable
2910 instead of to C<$_>:
2912 while (my $line = <STDIN>) { print $line }
2914 In these loop constructs, the assigned value (whether assignment
2915 is automatic or explicit) is then tested to see whether it is
2916 defined. The defined test avoids problems where the line has a string
2917 value that would be treated as false by Perl; for example a "" or
2918 a "0" with no trailing newline. If you really mean for such values
2919 to terminate the loop, they should be tested for explicitly:
2921 while (($_ = <STDIN>) ne '0') { ... }
2922 while (<STDIN>) { last unless $_; ... }
2924 In other boolean contexts, C<< <FILEHANDLE> >> without an
2925 explicit C<defined> test or comparison elicits a warning if the
2926 C<use warnings> pragma or the B<-w>
2927 command-line switch (the C<$^W> variable) is in effect.
2929 The filehandles STDIN, STDOUT, and STDERR are predefined. (The
2930 filehandles C<stdin>, C<stdout>, and C<stderr> will also work except
2931 in packages, where they would be interpreted as local identifiers
2932 rather than global.) Additional filehandles may be created with
2933 the open() function, amongst others. See L<perlopentut> and
2934 L<perlfunc/open> for details on this.
2935 X<stdin> X<stdout> X<sterr>
2937 If a <FILEHANDLE> is used in a context that is looking for
2938 a list, a list comprising all input lines is returned, one line per
2939 list element. It's easy to grow to a rather large data space this
2940 way, so use with care.
2942 <FILEHANDLE> may also be spelled C<readline(*FILEHANDLE)>.
2943 See L<perlfunc/readline>.
2945 The null filehandle <> is special: it can be used to emulate the
2946 behavior of B<sed> and B<awk>, and any other Unix filter program
2947 that takes a list of filenames, doing the same to each line
2948 of input from all of them. Input from <> comes either from
2949 standard input, or from each file listed on the command line. Here's
2950 how it works: the first time <> is evaluated, the @ARGV array is
2951 checked, and if it is empty, C<$ARGV[0]> is set to "-", which when opened
2952 gives you standard input. The @ARGV array is then processed as a list
2953 of filenames. The loop
2956 ... # code for each line
2959 is equivalent to the following Perl-like pseudo code:
2961 unshift(@ARGV, '-') unless @ARGV;
2962 while ($ARGV = shift) {
2965 ... # code for each line
2969 except that it isn't so cumbersome to say, and will actually work.
2970 It really does shift the @ARGV array and put the current filename
2971 into the $ARGV variable. It also uses filehandle I<ARGV>
2972 internally. <> is just a synonym for <ARGV>, which
2973 is magical. (The pseudo code above doesn't work because it treats
2974 <ARGV> as non-magical.)
2976 Since the null filehandle uses the two argument form of L<perlfunc/open>
2977 it interprets special characters, so if you have a script like this:
2983 and call it with C<perl dangerous.pl 'rm -rfv *|'>, it actually opens a
2984 pipe, executes the C<rm> command and reads C<rm>'s output from that pipe.
2985 If you want all items in C<@ARGV> to be interpreted as file names, you
2986 can use the module C<ARGV::readonly> from CPAN, or use the double bracket:
2992 Using double angle brackets inside of a while causes the open to use the
2993 three argument form (with the second argument being C<< < >>), so all
2994 arguments in ARGV are treated as literal filenames (including "-").
2995 (Note that for convenience, if you use C<< <<>> >> and if @ARGV is
2996 empty, it will still read from the standard input.)
2998 You can modify @ARGV before the first <> as long as the array ends up
2999 containing the list of filenames you really want. Line numbers (C<$.>)
3000 continue as though the input were one big happy file. See the example
3001 in L<perlfunc/eof> for how to reset line numbers on each file.
3003 If you want to set @ARGV to your own list of files, go right ahead.
3004 This sets @ARGV to all plain text files if no @ARGV was given:
3006 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
3008 You can even set them to pipe commands. For example, this automatically
3009 filters compressed arguments through B<gzip>:
3011 @ARGV = map { /\.(gz|Z)$/ ? "gzip -dc < $_ |" : $_ } @ARGV;
3013 If you want to pass switches into your script, you can use one of the
3014 Getopts modules or put a loop on the front like this:
3016 while ($_ = $ARGV[0], /^-/) {
3019 if (/^-D(.*)/) { $debug = $1 }
3020 if (/^-v/) { $verbose++ }
3021 # ... # other switches
3025 # ... # code for each line
3028 The <> symbol will return C<undef> for end-of-file only once.
3029 If you call it again after this, it will assume you are processing another
3030 @ARGV list, and if you haven't set @ARGV, will read input from STDIN.
3032 If what the angle brackets contain is a simple scalar variable (for example,
3033 <$foo>), then that variable contains the name of the
3034 filehandle to input from, or its typeglob, or a reference to the
3040 If what's within the angle brackets is neither a filehandle nor a simple
3041 scalar variable containing a filehandle name, typeglob, or typeglob
3042 reference, it is interpreted as a filename pattern to be globbed, and
3043 either a list of filenames or the next filename in the list is returned,
3044 depending on context. This distinction is determined on syntactic
3045 grounds alone. That means C<< <$x> >> is always a readline() from
3046 an indirect handle, but C<< <$hash{key}> >> is always a glob().
3047 That's because $x is a simple scalar variable, but C<$hash{key}> is
3048 not--it's a hash element. Even C<< <$x > >> (note the extra space)
3049 is treated as C<glob("$x ")>, not C<readline($x)>.
3051 One level of double-quote interpretation is done first, but you can't
3052 say C<< <$foo> >> because that's an indirect filehandle as explained
3053 in the previous paragraph. (In older versions of Perl, programmers
3054 would insert curly brackets to force interpretation as a filename glob:
3055 C<< <${foo}> >>. These days, it's considered cleaner to call the
3056 internal function directly as C<glob($foo)>, which is probably the right
3057 way to have done it in the first place.) For example:
3063 is roughly equivalent to:
3065 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
3071 except that the globbing is actually done internally using the standard
3072 C<File::Glob> extension. Of course, the shortest way to do the above is:
3076 A (file)glob evaluates its (embedded) argument only when it is
3077 starting a new list. All values must be read before it will start
3078 over. In list context, this isn't important because you automatically
3079 get them all anyway. However, in scalar context the operator returns
3080 the next value each time it's called, or C<undef> when the list has
3081 run out. As with filehandle reads, an automatic C<defined> is
3082 generated when the glob occurs in the test part of a C<while>,
3083 because legal glob returns (for example,
3084 a file called F<0>) would otherwise
3085 terminate the loop. Again, C<undef> is returned only once. So if
3086 you're expecting a single value from a glob, it is much better to
3089 ($file) = <blurch*>;
3095 because the latter will alternate between returning a filename and
3098 If you're trying to do variable interpolation, it's definitely better
3099 to use the glob() function, because the older notation can cause people
3100 to become confused with the indirect filehandle notation.
3102 @files = glob("$dir/*.[ch]");
3103 @files = glob($files[$i]);
3105 =head2 Constant Folding
3106 X<constant folding> X<folding>
3108 Like C, Perl does a certain amount of expression evaluation at
3109 compile time whenever it determines that all arguments to an
3110 operator are static and have no side effects. In particular, string
3111 concatenation happens at compile time between literals that don't do
3112 variable substitution. Backslash interpolation also happens at
3113 compile time. You can say
3115 'Now is the time for all'
3117 . 'good men to come to.'
3119 and this all reduces to one string internally. Likewise, if
3122 foreach $file (@filenames) {
3123 if (-s $file > 5 + 100 * 2**16) { }
3126 the compiler precomputes the number which that expression
3127 represents so that the interpreter won't have to.
3132 Perl doesn't officially have a no-op operator, but the bare constants
3133 C<0> and C<1> are special-cased not to produce a warning in void
3134 context, so you can for example safely do
3138 =head2 Bitwise String Operators
3139 X<operator, bitwise, string>
3141 Bitstrings of any size may be manipulated by the bitwise operators
3144 If the operands to a binary bitwise op are strings of different
3145 sizes, B<|> and B<^> ops act as though the shorter operand had
3146 additional zero bits on the right, while the B<&> op acts as though
3147 the longer operand were truncated to the length of the shorter.
3148 The granularity for such extension or truncation is one or more
3151 # ASCII-based examples
3152 print "j p \n" ^ " a h"; # prints "JAPH\n"
3153 print "JA" | " ph\n"; # prints "japh\n"
3154 print "japh\nJunk" & '_____'; # prints "JAPH\n";
3155 print 'p N$' ^ " E<H\n"; # prints "Perl\n";
3157 If you are intending to manipulate bitstrings, be certain that
3158 you're supplying bitstrings: If an operand is a number, that will imply
3159 a B<numeric> bitwise operation. You may explicitly show which type of
3160 operation you intend by using C<""> or C<0+>, as in the examples below.
3162 $foo = 150 | 105; # yields 255 (0x96 | 0x69 is 0xFF)
3163 $foo = '150' | 105; # yields 255
3164 $foo = 150 | '105'; # yields 255
3165 $foo = '150' | '105'; # yields string '155' (under ASCII)
3167 $baz = 0+$foo & 0+$bar; # both ops explicitly numeric
3168 $biz = "$foo" ^ "$bar"; # both ops explicitly stringy
3170 See L<perlfunc/vec> for information on how to manipulate individual bits
3173 =head2 Integer Arithmetic
3176 By default, Perl assumes that it must do most of its arithmetic in
3177 floating point. But by saying
3181 you may tell the compiler to use integer operations
3182 (see L<integer> for a detailed explanation) from here to the end of
3183 the enclosing BLOCK. An inner BLOCK may countermand this by saying
3187 which lasts until the end of that BLOCK. Note that this doesn't
3188 mean everything is an integer, merely that Perl will use integer
3189 operations for arithmetic, comparison, and bitwise operators. For
3190 example, even under C<use integer>, if you take the C<sqrt(2)>, you'll
3191 still get C<1.4142135623731> or so.
3193 Used on numbers, the bitwise operators ("&", "|", "^", "~", "<<",
3194 and ">>") always produce integral results. (But see also
3195 L<Bitwise String Operators>.) However, C<use integer> still has meaning for
3196 them. By default, their results are interpreted as unsigned integers, but
3197 if C<use integer> is in effect, their results are interpreted
3198 as signed integers. For example, C<~0> usually evaluates to a large
3199 integral value. However, C<use integer; ~0> is C<-1> on two's-complement
3202 =head2 Floating-point Arithmetic
3204 X<floating-point> X<floating point> X<float> X<real>
3206 While C<use integer> provides integer-only arithmetic, there is no
3207 analogous mechanism to provide automatic rounding or truncation to a
3208 certain number of decimal places. For rounding to a certain number
3209 of digits, sprintf() or printf() is usually the easiest route.
3212 Floating-point numbers are only approximations to what a mathematician
3213 would call real numbers. There are infinitely more reals than floats,
3214 so some corners must be cut. For example:
3216 printf "%.20g\n", 123456789123456789;
3217 # produces 123456789123456784
3219 Testing for exact floating-point equality or inequality is not a
3220 good idea. Here's a (relatively expensive) work-around to compare
3221 whether two floating-point numbers are equal to a particular number of
3222 decimal places. See Knuth, volume II, for a more robust treatment of
3226 my ($X, $Y, $POINTS) = @_;
3228 $tX = sprintf("%.${POINTS}g", $X);
3229 $tY = sprintf("%.${POINTS}g", $Y);
3233 The POSIX module (part of the standard perl distribution) implements
3234 ceil(), floor(), and other mathematical and trigonometric functions.
3235 The Math::Complex module (part of the standard perl distribution)
3236 defines mathematical functions that work on both the reals and the
3237 imaginary numbers. Math::Complex not as efficient as POSIX, but
3238 POSIX can't work with complex numbers.
3240 Rounding in financial applications can have serious implications, and
3241 the rounding method used should be specified precisely. In these
3242 cases, it probably pays not to trust whichever system rounding is
3243 being used by Perl, but to instead implement the rounding function you
3246 =head2 Bigger Numbers
3247 X<number, arbitrary precision>
3249 The standard C<Math::BigInt>, C<Math::BigRat>, and C<Math::BigFloat> modules,
3250 along with the C<bignum>, C<bigint>, and C<bigrat> pragmas, provide
3251 variable-precision arithmetic and overloaded operators, although
3252 they're currently pretty slow. At the cost of some space and
3253 considerable speed, they avoid the normal pitfalls associated with
3254 limited-precision representations.
3257 use bigint; # easy interface to Math::BigInt
3258 $x = 123456789123456789;
3260 +15241578780673678515622620750190521
3268 say "x/y is ", $x/$y;
3269 say "x*y is ", $x*$y;
3273 Several modules let you calculate with (bound only by memory and CPU time)
3274 unlimited or fixed precision. There
3275 are also some non-standard modules that
3276 provide faster implementations via external C libraries.
3278 Here is a short, but incomplete summary:
3280 Math::String treat string sequences like numbers
3281 Math::FixedPrecision calculate with a fixed precision
3282 Math::Currency for currency calculations
3283 Bit::Vector manipulate bit vectors fast (uses C)
3284 Math::BigIntFast Bit::Vector wrapper for big numbers
3285 Math::Pari provides access to the Pari C library
3286 Math::Cephes uses the external Cephes C library (no
3288 Math::Cephes::Fraction fractions via the Cephes library
3289 Math::GMP another one using an external C library
3290 Math::GMPz an alternative interface to libgmp's big ints
3291 Math::GMPq an interface to libgmp's fraction numbers
3292 Math::GMPf an interface to libgmp's floating point numbers