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 S<C<$x + $y>>
10 is always a numeric addition, and if C<$x> or C<$y> do not contain
11 numbers, an attempt is made to convert them to numbers first.
13 This is in contrast to many other dynamic languages, where the
14 operation is determined by the type of the first argument. It also
15 means that Perl has two versions of some operators, one for numeric
16 and one for string comparison. For example S<C<$x == $y>> compares
17 two numbers for equality, and S<C<$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<|>, 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 group more tightly than others.
31 For example, in C<2 + 4 * 5>, the multiplication has higher precedence, so C<4
32 * 5> is grouped together as the right-hand operand of the addition, rather
33 than C<2 + 4> being grouped together as the left-hand operand of the
34 multiplication. It is as if the expression were written C<2 + (4 * 5)>, not
35 C<(2 + 4) * 5>. So the expression yields C<2 + 20 == 22>, rather than
38 I<Operator associativity> defines what happens if a sequence of the same
39 operators is used one after another:
40 usually that they will be grouped at the left
41 or the right. For example, in C<9 - 3 - 2>, subtraction is left associative,
42 so C<9 - 3> is grouped together as the left-hand operand of the second
43 subtraction, rather than C<3 - 2> being grouped together as the right-hand
44 operand of the first subtraction. It is as if the expression were written
45 C<(9 - 3) - 2>, not C<9 - (3 - 2)>. So the expression yields C<6 - 2 == 4>,
46 rather than C<9 - 1 == 8>.
48 For simple operators that evaluate all their operands and then combine the
49 values in some way, precedence and associativity (and parentheses) imply some
50 ordering requirements on those combining operations. For example, in C<2 + 4 *
51 5>, the grouping implied by precedence means that the multiplication of 4 and
52 5 must be performed before the addition of 2 and 20, simply because the result
53 of that multiplication is required as one of the operands of the addition. But
54 the order of operations is not fully determined by this: in C<2 * 2 + 4 * 5>
55 both multiplications must be performed before the addition, but the grouping
56 does not say anything about the order in which the two multiplications are
57 performed. In fact Perl has a general rule that the operands of an operator
58 are evaluated in left-to-right order. A few operators such as C<&&=> have
59 special evaluation rules that can result in an operand not being evaluated at
60 all; in general, the top-level operator in an expression has control of
63 Some comparison operators, as their associativity, I<chain> with some
64 operators of the same precedence (but never with operators of different
65 precedence). This chaining means that each comparison is performed
66 on the two arguments surrounding it, with each interior argument taking
67 part in two comparisons, and the comparison results are implicitly ANDed.
68 Thus S<C<"$x E<lt> $y E<lt>= $z">> behaves exactly like S<C<"$x E<lt>
69 $y && $y E<lt>= $z">>, assuming that C<"$y"> is as simple a scalar as
70 it looks. The ANDing short-circuits just like C<"&&"> does, stopping
71 the sequence of comparisons as soon as one yields false.
73 In a chained comparison, each argument expression is evaluated at most
74 once, even if it takes part in two comparisons, but the result of the
75 evaluation is fetched for each comparison. (It is not evaluated
76 at all if the short-circuiting means that it's not required for any
77 comparisons.) This matters if the computation of an interior argument
78 is expensive or non-deterministic. For example,
80 if($x < expensive_sub() <= $z) { ...
84 if($x < expensive_sub() && expensive_sub() <= $z) { ...
88 my $tmp = expensive_sub();
89 if($x < $tmp && $tmp <= $z) { ...
91 in that the subroutine is only called once. However, it's not exactly
92 like this latter code either, because the chained comparison doesn't
93 actually involve any temporary variable (named or otherwise): there is
94 no assignment. This doesn't make much difference where the expression
95 is a call to an ordinary subroutine, but matters more with an lvalue
96 subroutine, or if the argument expression yields some unusual kind of
97 scalar by other means. For example, if the argument expression yields
98 a tied scalar, then the expression is evaluated to produce that scalar
99 at most once, but the value of that scalar may be fetched up to twice,
100 once for each comparison in which it is actually used.
102 In this example, the expression is evaluated only once, and the tied
103 scalar (the result of the expression) is fetched for each comparison that
106 if ($x < $tied_scalar < $z) { ...
108 In the next example, the expression is evaluated only once, and the tied
109 scalar is fetched once as part of the operation within the expression.
110 The result of that operation is fetched for each comparison, which
111 normally doesn't matter unless that expression result is also magical due
112 to operator overloading.
114 if ($x < $tied_scalar + 42 < $z) { ...
116 Some operators are instead non-associative, meaning that it is a syntax
117 error to use a sequence of those operators of the same precedence.
118 For example, S<C<"$x .. $y .. $z">> is an error.
120 Perl operators have the following associativity and precedence,
121 listed from highest precedence to lowest. Operators borrowed from
122 C keep the same precedence relationship with each other, even where
123 C's precedence is slightly screwy. (This makes learning Perl easier
124 for C folks.) With very few exceptions, these all operate on scalar
125 values only, not array values.
127 left terms and list operators (leftward)
131 right ! ~ ~. \ and unary + and -
136 nonassoc named unary operators
138 chained < > <= >= lt gt le ge
139 chain/na == != eq ne <=> cmp ~~
146 right = += -= *= etc. goto last next redo dump
148 nonassoc list operators (rightward)
153 In the following sections, these operators are covered in detail, in the
154 same order in which they appear in the table above.
156 Many operators can be overloaded for objects. See L<overload>.
158 =head2 Terms and List Operators (Leftward)
159 X<list operator> X<operator, list> X<term>
161 A TERM has the highest precedence in Perl. They include variables,
162 quote and quote-like operators, any expression in parentheses,
163 and any function whose arguments are parenthesized. Actually, there
164 aren't really functions in this sense, just list operators and unary
165 operators behaving as functions because you put parentheses around
166 the arguments. These are all documented in L<perlfunc>.
168 If any list operator (C<print()>, etc.) or any unary operator (C<chdir()>, etc.)
169 is followed by a left parenthesis as the next token, the operator and
170 arguments within parentheses are taken to be of highest precedence,
171 just like a normal function call.
173 In the absence of parentheses, the precedence of list operators such as
174 C<print>, C<sort>, or C<chmod> is either very high or very low depending on
175 whether you are looking at the left side or the right side of the operator.
178 @ary = (1, 3, sort 4, 2);
179 print @ary; # prints 1324
181 the commas on the right of the C<sort> are evaluated before the C<sort>,
182 but the commas on the left are evaluated after. In other words,
183 list operators tend to gobble up all arguments that follow, and
184 then act like a simple TERM with regard to the preceding expression.
185 Be careful with parentheses:
187 # These evaluate exit before doing the print:
188 print($foo, exit); # Obviously not what you want.
189 print $foo, exit; # Nor is this.
191 # These do the print before evaluating exit:
192 (print $foo), exit; # This is what you want.
193 print($foo), exit; # Or this.
194 print ($foo), exit; # Or even this.
198 print ($foo & 255) + 1, "\n";
200 probably doesn't do what you expect at first glance. The parentheses
201 enclose the argument list for C<print> which is evaluated (printing
202 the result of S<C<$foo & 255>>). Then one is added to the return value
203 of C<print> (usually 1). The result is something like this:
205 1 + 1, "\n"; # Obviously not what you meant.
207 To do what you meant properly, you must write:
209 print(($foo & 255) + 1, "\n");
211 See L</Named Unary Operators> for more discussion of this.
213 Also parsed as terms are the S<C<do {}>> and S<C<eval {}>> constructs, as
214 well as subroutine and method calls, and the anonymous
215 constructors C<[]> and C<{}>.
217 See also L</Quote and Quote-like Operators> toward the end of this section,
218 as well as L</"I/O Operators">.
220 =head2 The Arrow Operator
221 X<arrow> X<dereference> X<< -> >>
223 "C<< -> >>" is an infix dereference operator, just as it is in C
224 and C++. If the right side is either a C<[...]>, C<{...}>, or a
225 C<(...)> subscript, then the left side must be either a hard or
226 symbolic reference to an array, a hash, or a subroutine respectively.
227 (Or technically speaking, a location capable of holding a hard
228 reference, if it's an array or hash reference being used for
229 assignment.) See L<perlreftut> and L<perlref>.
231 Otherwise, the right side is a method name or a simple scalar
232 variable containing either the method name or a subroutine reference,
233 and (if it is a method name) the left side must be either an object (a
234 blessed reference) or a class name (that is, a package name). See
237 The dereferencing cases (as opposed to method-calling cases) are
238 somewhat extended by the C<postderef> feature. For the
239 details of that feature, consult L<perlref/Postfix Dereference Syntax>.
241 =head2 Auto-increment and Auto-decrement
242 X<increment> X<auto-increment> X<++> X<decrement> X<auto-decrement> X<-->
244 C<"++"> and C<"--"> work as in C. That is, if placed before a variable,
245 they increment or decrement the variable by one before returning the
246 value, and if placed after, increment or decrement after returning the
250 print $i++; # prints 0
251 print ++$j; # prints 1
253 Note that just as in C, Perl doesn't define B<when> the variable is
254 incremented or decremented. You just know it will be done sometime
255 before or after the value is returned. This also means that modifying
256 a variable twice in the same statement will lead to undefined behavior.
257 Avoid statements like:
262 Perl will not guarantee what the result of the above statements is.
264 The auto-increment operator has a little extra builtin magic to it. If
265 you increment a variable that is numeric, or that has ever been used in
266 a numeric context, you get a normal increment. If, however, the
267 variable has been used in only string contexts since it was set, and
268 has a value that is not the empty string and matches the pattern
269 C</^[a-zA-Z]*[0-9]*\z/>, the increment is done as a string, preserving each
270 character within its range, with carry:
272 print ++($foo = "99"); # prints "100"
273 print ++($foo = "a0"); # prints "a1"
274 print ++($foo = "Az"); # prints "Ba"
275 print ++($foo = "zz"); # prints "aaa"
277 C<undef> is always treated as numeric, and in particular is changed
278 to C<0> before incrementing (so that a post-increment of an undef value
279 will return C<0> rather than C<undef>).
281 The auto-decrement operator is not magical.
283 =head2 Exponentiation
284 X<**> X<exponentiation> X<power>
286 Binary C<"**"> is the exponentiation operator. It binds even more
287 tightly than unary minus, so C<-2**4> is C<-(2**4)>, not C<(-2)**4>.
289 implemented using C's C<pow(3)> function, which actually works on doubles
292 Note that certain exponentiation expressions are ill-defined:
293 these include C<0**0>, C<1**Inf>, and C<Inf**0>. Do not expect
294 any particular results from these special cases, the results
295 are platform-dependent.
297 =head2 Symbolic Unary Operators
298 X<unary operator> X<operator, unary>
300 Unary C<"!"> performs logical negation, that is, "not". See also
301 L<C<not>|/Logical Not> for a lower precedence version of this.
304 Unary C<"-"> performs arithmetic negation if the operand is numeric,
305 including any string that looks like a number. If the operand is
306 an identifier, a string consisting of a minus sign concatenated
307 with the identifier is returned. Otherwise, if the string starts
308 with a plus or minus, a string starting with the opposite sign is
309 returned. One effect of these rules is that C<-bareword> is equivalent
310 to the string C<"-bareword">. If, however, the string begins with a
311 non-alphabetic character (excluding C<"+"> or C<"-">), Perl will attempt
313 the string to a numeric, and the arithmetic negation is performed. If the
314 string cannot be cleanly converted to a numeric, Perl will give the warning
315 B<Argument "the string" isn't numeric in negation (-) at ...>.
316 X<-> X<negation, arithmetic>
318 Unary C<"~"> performs bitwise negation, that is, 1's complement. For
319 example, S<C<0666 & ~027>> is 0640. (See also L</Integer Arithmetic> and
320 L</Bitwise String Operators>.) Note that the width of the result is
321 platform-dependent: C<~0> is 32 bits wide on a 32-bit platform, but 64
322 bits wide on a 64-bit platform, so if you are expecting a certain bit
323 width, remember to use the C<"&"> operator to mask off the excess bits.
324 X<~> X<negation, binary>
326 Starting in Perl 5.28, it is a fatal error to try to complement a string
327 containing a character with an ordinal value above 255.
329 If the "bitwise" feature is enabled via S<C<use
330 feature 'bitwise'>> or C<use v5.28>, then unary
331 C<"~"> always treats its argument as a number, and an
332 alternate form of the operator, C<"~.">, always treats its argument as a
333 string. So C<~0> and C<~"0"> will both give 2**32-1 on 32-bit platforms,
334 whereas C<~.0> and C<~."0"> will both yield C<"\xff">. Until Perl 5.28,
335 this feature produced a warning in the C<"experimental::bitwise"> category.
337 Unary C<"+"> has no effect whatsoever, even on strings. It is useful
338 syntactically for separating a function name from a parenthesized expression
339 that would otherwise be interpreted as the complete list of function
340 arguments. (See examples above under L</Terms and List Operators (Leftward)>.)
343 Unary C<"\"> creates references. If its operand is a single sigilled
344 thing, it creates a reference to that object. If its operand is a
345 parenthesised list, then it creates references to the things mentioned
346 in the list. Otherwise it puts its operand in list context, and creates
347 a list of references to the scalars in the list provided by the operand.
349 and L<perlref>. Do not confuse this behavior with the behavior of
350 backslash within a string, although both forms do convey the notion
351 of protecting the next thing from interpolation.
352 X<\> X<reference> X<backslash>
354 =head2 Binding Operators
355 X<binding> X<operator, binding> X<=~> X<!~>
357 Binary C<"=~"> binds a scalar expression to a pattern match. Certain operations
358 search or modify the string C<$_> by default. This operator makes that kind
359 of operation work on some other string. The right argument is a search
360 pattern, substitution, or transliteration. The left argument is what is
361 supposed to be searched, substituted, or transliterated instead of the default
362 C<$_>. When used in scalar context, the return value generally indicates the
363 success of the operation. The exceptions are substitution (C<s///>)
364 and transliteration (C<y///>) with the C</r> (non-destructive) option,
365 which cause the B<r>eturn value to be the result of the substitution.
366 Behavior in list context depends on the particular operator.
367 See L</"Regexp Quote-Like Operators"> for details and L<perlretut> for
368 examples using these operators.
370 If the right argument is an expression rather than a search pattern,
371 substitution, or transliteration, it is interpreted as a search pattern at run
372 time. Note that this means that its
373 contents will be interpolated twice, so
377 is not ok, as the regex engine will end up trying to compile the
378 pattern C<\>, which it will consider a syntax error.
380 Binary C<"!~"> is just like C<"=~"> except the return value is negated in
383 Binary C<"!~"> with a non-destructive substitution (C<s///r>) or transliteration
384 (C<y///r>) is a syntax error.
386 =head2 Multiplicative Operators
387 X<operator, multiplicative>
389 Binary C<"*"> multiplies two numbers.
392 Binary C<"/"> divides two numbers.
395 Binary C<"%"> is the modulo operator, which computes the division
396 remainder of its first argument with respect to its second argument.
398 operands C<$m> and C<$n>: If C<$n> is positive, then S<C<$m % $n>> is
399 C<$m> minus the largest multiple of C<$n> less than or equal to
400 C<$m>. If C<$n> is negative, then S<C<$m % $n>> is C<$m> minus the
401 smallest multiple of C<$n> that is not less than C<$m> (that is, the
402 result will be less than or equal to zero). If the operands
403 C<$m> and C<$n> are floating point values and the absolute value of
404 C<$n> (that is C<abs($n)>) is less than S<C<(UV_MAX + 1)>>, only
405 the integer portion of C<$m> and C<$n> will be used in the operation
406 (Note: here C<UV_MAX> means the maximum of the unsigned integer type).
407 If the absolute value of the right operand (C<abs($n)>) is greater than
408 or equal to S<C<(UV_MAX + 1)>>, C<"%"> computes the floating-point remainder
409 C<$r> in the equation S<C<($r = $m - $i*$n)>> where C<$i> is a certain
410 integer that makes C<$r> have the same sign as the right operand
411 C<$n> (B<not> as the left operand C<$m> like C function C<fmod()>)
412 and the absolute value less than that of C<$n>.
413 Note that when S<C<use integer>> is in scope, C<"%"> gives you direct access
414 to the modulo operator as implemented by your C compiler. This
415 operator is not as well defined for negative operands, but it will
417 X<%> X<remainder> X<modulo> X<mod>
419 Binary C<x> is the repetition operator. In scalar context, or if the
420 left operand is neither enclosed in parentheses nor a C<qw//> list,
421 it performs a string repetition. In that case it supplies scalar
422 context to the left operand, and returns a string consisting of the
423 left operand string repeated the number of times specified by the right
424 operand. If the C<x> is in list context, and the left operand is either
425 enclosed in parentheses or a C<qw//> list, it performs a list repetition.
426 In that case it supplies list context to the left operand, and returns
427 a list consisting of the left operand list repeated the number of times
428 specified by the right operand.
429 If the right operand is zero or negative (raising a warning on
430 negative), it returns an empty string
431 or an empty list, depending on the context.
434 print '-' x 80; # print row of dashes
436 print "\t" x ($tab/8), ' ' x ($tab%8); # tab over
438 @ones = (1) x 80; # a list of 80 1's
439 @ones = (5) x @ones; # set all elements to 5
442 =head2 Additive Operators
443 X<operator, additive>
445 Binary C<"+"> returns the sum of two numbers.
448 Binary C<"-"> returns the difference of two numbers.
451 Binary C<"."> concatenates two strings.
452 X<string, concatenation> X<concatenation>
453 X<cat> X<concat> X<concatenate> X<.>
455 =head2 Shift Operators
456 X<shift operator> X<operator, shift> X<<< << >>>
457 X<<< >> >>> X<right shift> X<left shift> X<bitwise shift>
458 X<shl> X<shr> X<shift, right> X<shift, left>
460 Binary C<<< "<<" >>> returns the value of its left argument shifted left by the
461 number of bits specified by the right argument. Arguments should be
462 integers. (See also L</Integer Arithmetic>.)
464 Binary C<<< ">>" >>> returns the value of its left argument shifted right by
465 the number of bits specified by the right argument. Arguments should
466 be integers. (See also L</Integer Arithmetic>.)
468 If S<C<use integer>> (see L</Integer Arithmetic>) is in force then
469 signed C integers are used (I<arithmetic shift>), otherwise unsigned C
470 integers are used (I<logical shift>), even for negative shiftees.
471 In arithmetic right shift the sign bit is replicated on the left,
472 in logical shift zero bits come in from the left.
474 Either way, the implementation isn't going to generate results larger
475 than the size of the integer type Perl was built with (32 bits or 64 bits).
477 Shifting by negative number of bits means the reverse shift: left
478 shift becomes right shift, right shift becomes left shift. This is
479 unlike in C, where negative shift is undefined.
481 Shifting by more bits than the size of the integers means most of the
482 time zero (all bits fall off), except that under S<C<use integer>>
483 right overshifting a negative shiftee results in -1. This is unlike
484 in C, where shifting by too many bits is undefined. A common C
485 behavior is "shift by modulo wordbits", so that for example
487 1 >> 64 == 1 >> (64 % 64) == 1 >> 0 == 1 # Common C behavior.
489 but that is completely accidental.
491 If you get tired of being subject to your platform's native integers,
492 the S<C<use bigint>> pragma neatly sidesteps the issue altogether:
494 print 20 << 20; # 20971520
495 print 20 << 40; # 5120 on 32-bit machines,
496 # 21990232555520 on 64-bit machines
498 print 20 << 100; # 25353012004564588029934064107520
500 =head2 Named Unary Operators
501 X<operator, named unary>
503 The various named unary operators are treated as functions with one
504 argument, with optional parentheses.
506 If any list operator (C<print()>, etc.) or any unary operator (C<chdir()>, etc.)
507 is followed by a left parenthesis as the next token, the operator and
508 arguments within parentheses are taken to be of highest precedence,
509 just like a normal function call. For example,
510 because named unary operators are higher precedence than C<||>:
512 chdir $foo || die; # (chdir $foo) || die
513 chdir($foo) || die; # (chdir $foo) || die
514 chdir ($foo) || die; # (chdir $foo) || die
515 chdir +($foo) || die; # (chdir $foo) || die
517 but, because C<"*"> is higher precedence than named operators:
519 chdir $foo * 20; # chdir ($foo * 20)
520 chdir($foo) * 20; # (chdir $foo) * 20
521 chdir ($foo) * 20; # (chdir $foo) * 20
522 chdir +($foo) * 20; # chdir ($foo * 20)
524 rand 10 * 20; # rand (10 * 20)
525 rand(10) * 20; # (rand 10) * 20
526 rand (10) * 20; # (rand 10) * 20
527 rand +(10) * 20; # rand (10 * 20)
529 Regarding precedence, the filetest operators, like C<-f>, C<-M>, etc. are
530 treated like named unary operators, but they don't follow this functional
531 parenthesis rule. That means, for example, that C<-f($file).".bak"> is
532 equivalent to S<C<-f "$file.bak">>.
533 X<-X> X<filetest> X<operator, filetest>
535 See also L</"Terms and List Operators (Leftward)">.
537 =head2 Relational Operators
538 X<relational operator> X<operator, relational>
540 Perl operators that return true or false generally return values
541 that can be safely used as numbers. For example, the relational
542 operators in this section and the equality operators in the next
543 one return C<1> for true and a special version of the defined empty
544 string, C<"">, which counts as a zero but is exempt from warnings
545 about improper numeric conversions, just as S<C<"0 but true">> is.
547 Binary C<< "<" >> returns true if the left argument is numerically less than
551 Binary C<< ">" >> returns true if the left argument is numerically greater
552 than the right argument.
555 Binary C<< "<=" >> returns true if the left argument is numerically less than
556 or equal to the right argument.
559 Binary C<< ">=" >> returns true if the left argument is numerically greater
560 than or equal to the right argument.
563 Binary C<"lt"> returns true if the left argument is stringwise less than
567 Binary C<"gt"> returns true if the left argument is stringwise greater
568 than the right argument.
571 Binary C<"le"> returns true if the left argument is stringwise less than
572 or equal to the right argument.
575 Binary C<"ge"> returns true if the left argument is stringwise greater
576 than or equal to the right argument.
579 A sequence of relational operators, such as S<C<"$x E<lt> $y E<lt>=
580 $z">>, performs chained comparisons, in the manner described above in
581 the section L</"Operator Precedence and Associativity">.
582 Beware that they do not chain with equality operators, which have lower
585 =head2 Equality Operators
586 X<equality> X<equal> X<equals> X<operator, equality>
588 Binary C<< "==" >> returns true if the left argument is numerically equal to
592 Binary C<< "!=" >> returns true if the left argument is numerically not equal
593 to the right argument.
596 Binary C<"eq"> returns true if the left argument is stringwise equal to
600 Binary C<"ne"> returns true if the left argument is stringwise not equal
601 to the right argument.
604 A sequence of the above equality operators, such as S<C<"$x == $y ==
605 $z">>, performs chained comparisons, in the manner described above in
606 the section L</"Operator Precedence and Associativity">.
607 Beware that they do not chain with relational operators, which have
610 Binary C<< "<=>" >> returns -1, 0, or 1 depending on whether the left
611 argument is numerically less than, equal to, or greater than the right
612 argument. If your platform supports C<NaN>'s (not-a-numbers) as numeric
613 values, using them with C<< "<=>" >> returns undef. C<NaN> is not
614 C<< "<" >>, C<< "==" >>, C<< ">" >>, C<< "<=" >> or C<< ">=" >> anything
615 (even C<NaN>), so those 5 return false. S<C<< NaN != NaN >>> returns
616 true, as does S<C<NaN !=> I<anything else>>. If your platform doesn't
617 support C<NaN>'s then C<NaN> is just a string with numeric value 0.
621 $ perl -le '$x = "NaN"; print "No NaN support here" if $x == $x'
622 $ perl -le '$x = "NaN"; print "NaN support here" if $x != $x'
624 (Note that the L<bigint>, L<bigrat>, and L<bignum> pragmas all
627 Binary C<"cmp"> returns -1, 0, or 1 depending on whether the left
628 argument is stringwise less than, equal to, or greater than the right
631 Here we can see the difference between <=> and cmp,
633 print 10 <=> 2 #prints 1
634 print 10 cmp 2 #prints -1
636 (likewise between gt and >, lt and <, etc.)
639 Binary C<"~~"> does a smartmatch between its arguments. Smart matching
640 is described in the next section.
643 The two-sided ordering operators C<"E<lt>=E<gt>"> and C<"cmp">, and the
644 smartmatch operator C<"~~">, are non-associative with respect to each
645 other and with respect to the equality operators of the same precedence.
647 C<"lt">, C<"le">, C<"ge">, C<"gt"> and C<"cmp"> use the collation (sort)
648 order specified by the current C<LC_COLLATE> locale if a S<C<use
649 locale>> form that includes collation is in effect. See L<perllocale>.
650 Do not mix these with Unicode,
651 only use them with legacy 8-bit locale encodings.
652 The standard C<L<Unicode::Collate>> and
653 C<L<Unicode::Collate::Locale>> modules offer much more powerful
654 solutions to collation issues.
656 For case-insensitive comparisons, look at the L<perlfunc/fc> case-folding
657 function, available in Perl v5.16 or later:
659 if ( fc($x) eq fc($y) ) { ... }
661 =head2 Class Instance Operator
664 Binary C<isa> evaluates to true when the left argument is an object instance of
665 the class (or a subclass derived from that class) given by the right argument.
666 If the left argument is not defined, not a blessed object instance, nor does
667 not derive from the class given by the right argument, the operator evaluates
668 as false. The right argument may give the class either as a bareword or a
669 scalar expression that yields a string class name:
671 if( $obj isa Some::Class ) { ... }
673 if( $obj isa "Different::Class" ) { ... }
674 if( $obj isa $name_of_class ) { ... }
676 This is an experimental feature and is available from Perl 5.31.6 when enabled
677 by C<use feature 'isa'>. It emits a warning in the C<experimental::isa>
680 =head2 Smartmatch Operator
682 First available in Perl 5.10.1 (the 5.10.0 version behaved differently),
683 binary C<~~> does a "smartmatch" between its arguments. This is mostly
684 used implicitly in the C<when> construct described in L<perlsyn>, although
685 not all C<when> clauses call the smartmatch operator. Unique among all of
686 Perl's operators, the smartmatch operator can recurse. The smartmatch
687 operator is L<experimental|perlpolicy/experimental> and its behavior is
690 It is also unique in that all other Perl operators impose a context
691 (usually string or numeric context) on their operands, autoconverting
692 those operands to those imposed contexts. In contrast, smartmatch
693 I<infers> contexts from the actual types of its operands and uses that
694 type information to select a suitable comparison mechanism.
696 The C<~~> operator compares its operands "polymorphically", determining how
697 to compare them according to their actual types (numeric, string, array,
698 hash, etc.). Like the equality operators with which it shares the same
699 precedence, C<~~> returns 1 for true and C<""> for false. It is often best
700 read aloud as "in", "inside of", or "is contained in", because the left
701 operand is often looked for I<inside> the right operand. That makes the
702 order of the operands to the smartmatch operand often opposite that of
703 the regular match operator. In other words, the "smaller" thing is usually
704 placed in the left operand and the larger one in the right.
706 The behavior of a smartmatch depends on what type of things its arguments
707 are, as determined by the following table. The first row of the table
708 whose types apply determines the smartmatch behavior. Because what
709 actually happens is mostly determined by the type of the second operand,
710 the table is sorted on the right operand instead of on the left.
712 Left Right Description and pseudocode
713 ===============================================================
714 Any undef check whether Any is undefined
717 Any Object invoke ~~ overloading on Object, or die
719 Right operand is an ARRAY:
721 Left Right Description and pseudocode
722 ===============================================================
723 ARRAY1 ARRAY2 recurse on paired elements of ARRAY1 and ARRAY2[2]
724 like: (ARRAY1[0] ~~ ARRAY2[0])
725 && (ARRAY1[1] ~~ ARRAY2[1]) && ...
726 HASH ARRAY any ARRAY elements exist as HASH keys
727 like: grep { exists HASH->{$_} } ARRAY
728 Regexp ARRAY any ARRAY elements pattern match Regexp
729 like: grep { /Regexp/ } ARRAY
730 undef ARRAY undef in ARRAY
731 like: grep { !defined } ARRAY
732 Any ARRAY smartmatch each ARRAY element[3]
733 like: grep { Any ~~ $_ } ARRAY
735 Right operand is a HASH:
737 Left Right Description and pseudocode
738 ===============================================================
739 HASH1 HASH2 all same keys in both HASHes
741 grep { exists HASH2->{$_} } keys HASH1
742 ARRAY HASH any ARRAY elements exist as HASH keys
743 like: grep { exists HASH->{$_} } ARRAY
744 Regexp HASH any HASH keys pattern match Regexp
745 like: grep { /Regexp/ } keys HASH
746 undef HASH always false (undef can't be a key)
748 Any HASH HASH key existence
749 like: exists HASH->{Any}
751 Right operand is CODE:
753 Left Right Description and pseudocode
754 ===============================================================
755 ARRAY CODE sub returns true on all ARRAY elements[1]
756 like: !grep { !CODE->($_) } ARRAY
757 HASH CODE sub returns true on all HASH keys[1]
758 like: !grep { !CODE->($_) } keys HASH
759 Any CODE sub passed Any returns true
762 Right operand is a Regexp:
764 Left Right Description and pseudocode
765 ===============================================================
766 ARRAY Regexp any ARRAY elements match Regexp
767 like: grep { /Regexp/ } ARRAY
768 HASH Regexp any HASH keys match Regexp
769 like: grep { /Regexp/ } keys HASH
770 Any Regexp pattern match
771 like: Any =~ /Regexp/
775 Left Right Description and pseudocode
776 ===============================================================
777 Object Any invoke ~~ overloading on Object,
780 Any Num numeric equality
782 Num nummy[4] numeric equality
784 undef Any check whether undefined
786 Any Any string equality
795 Empty hashes or arrays match.
798 That is, each element smartmatches the element of the same index in the other array.[3]
801 If a circular reference is found, fall back to referential equality.
804 Either an actual number, or a string that looks like one.
808 The smartmatch implicitly dereferences any non-blessed hash or array
809 reference, so the C<I<HASH>> and C<I<ARRAY>> entries apply in those cases.
810 For blessed references, the C<I<Object>> entries apply. Smartmatches
811 involving hashes only consider hash keys, never hash values.
813 The "like" code entry is not always an exact rendition. For example, the
814 smartmatch operator short-circuits whenever possible, but C<grep> does
815 not. Also, C<grep> in scalar context returns the number of matches, but
816 C<~~> returns only true or false.
818 Unlike most operators, the smartmatch operator knows to treat C<undef>
822 @array = (1, 2, 3, undef, 4, 5);
823 say "some elements undefined" if undef ~~ @array;
825 Each operand is considered in a modified scalar context, the modification
826 being that array and hash variables are passed by reference to the
827 operator, which implicitly dereferences them. Both elements
828 of each pair are the same:
832 my %hash = (red => 1, blue => 2, green => 3,
833 orange => 4, yellow => 5, purple => 6,
834 black => 7, grey => 8, white => 9);
836 my @array = qw(red blue green);
838 say "some array elements in hash keys" if @array ~~ %hash;
839 say "some array elements in hash keys" if \@array ~~ \%hash;
841 say "red in array" if "red" ~~ @array;
842 say "red in array" if "red" ~~ \@array;
844 say "some keys end in e" if /e$/ ~~ %hash;
845 say "some keys end in e" if /e$/ ~~ \%hash;
847 Two arrays smartmatch if each element in the first array smartmatches
848 (that is, is "in") the corresponding element in the second array,
852 my @little = qw(red blue green);
853 my @bigger = ("red", "blue", [ "orange", "green" ] );
854 if (@little ~~ @bigger) { # true!
855 say "little is contained in bigger";
858 Because the smartmatch operator recurses on nested arrays, this
859 will still report that "red" is in the array.
862 my @array = qw(red blue green);
863 my $nested_array = [[[[[[[ @array ]]]]]]];
864 say "red in array" if "red" ~~ $nested_array;
866 If two arrays smartmatch each other, then they are deep
867 copies of each others' values, as this example reports:
870 my @a = (0, 1, 2, [3, [4, 5], 6], 7);
871 my @b = (0, 1, 2, [3, [4, 5], 6], 7);
873 if (@a ~~ @b && @b ~~ @a) {
874 say "a and b are deep copies of each other";
877 say "a smartmatches in b";
880 say "b smartmatches in a";
883 say "a and b don't smartmatch each other at all";
887 If you were to set S<C<$b[3] = 4>>, then instead of reporting that "a and b
888 are deep copies of each other", it now reports that C<"b smartmatches in a">.
889 That's because the corresponding position in C<@a> contains an array that
890 (eventually) has a 4 in it.
892 Smartmatching one hash against another reports whether both contain the
893 same keys, no more and no less. This could be used to see whether two
894 records have the same field names, without caring what values those fields
895 might have. For example:
899 state $REQUIRED_FIELDS = { name=>1, rank=>1, serial_num=>1 };
901 my ($class, $init_fields) = @_;
903 die "Must supply (only) name, rank, and serial number"
904 unless $init_fields ~~ $REQUIRED_FIELDS;
909 However, this only does what you mean if C<$init_fields> is indeed a hash
910 reference. The condition C<$init_fields ~~ $REQUIRED_FIELDS> also allows the
911 strings C<"name">, C<"rank">, C<"serial_num"> as well as any array reference
912 that contains C<"name"> or C<"rank"> or C<"serial_num"> anywhere to pass
915 The smartmatch operator is most often used as the implicit operator of a
916 C<when> clause. See the section on "Switch Statements" in L<perlsyn>.
918 =head3 Smartmatching of Objects
920 To avoid relying on an object's underlying representation, if the
921 smartmatch's right operand is an object that doesn't overload C<~~>,
922 it raises the exception "C<Smartmatching a non-overloaded object
923 breaks encapsulation>". That's because one has no business digging
924 around to see whether something is "in" an object. These are all
925 illegal on objects without a C<~~> overload:
931 However, you can change the way an object is smartmatched by overloading
932 the C<~~> operator. This is allowed to
933 extend the usual smartmatch semantics.
934 For objects that do have an C<~~> overload, see L<overload>.
936 Using an object as the left operand is allowed, although not very useful.
937 Smartmatching rules take precedence over overloading, so even if the
938 object in the left operand has smartmatch overloading, this will be
939 ignored. A left operand that is a non-overloaded object falls back on a
940 string or numeric comparison of whatever the C<ref> operator returns. That
945 does I<not> invoke the overload method with C<I<X>> as an argument.
946 Instead the above table is consulted as normal, and based on the type of
947 C<I<X>>, overloading may or may not be invoked. For simple strings or
948 numbers, "in" becomes equivalent to this:
950 $object ~~ $number ref($object) == $number
951 $object ~~ $string ref($object) eq $string
953 For example, this reports that the handle smells IOish
954 (but please don't really do this!):
957 my $fh = IO::Handle->new();
958 if ($fh ~~ /\bIO\b/) {
959 say "handle smells IOish";
962 That's because it treats C<$fh> as a string like
963 C<"IO::Handle=GLOB(0x8039e0)">, then pattern matches against that.
966 X<operator, bitwise, and> X<bitwise and> X<&>
968 Binary C<"&"> returns its operands ANDed together bit by bit. Although no
969 warning is currently raised, the result is not well defined when this operation
970 is performed on operands that aren't either numbers (see
971 L</Integer Arithmetic>) nor bitstrings (see L</Bitwise String Operators>).
973 Note that C<"&"> has lower priority than relational operators, so for example
974 the parentheses are essential in a test like
976 print "Even\n" if ($x & 1) == 0;
978 If the "bitwise" feature is enabled via S<C<use feature 'bitwise'>> or
979 C<use v5.28>, then this operator always treats its operands as numbers.
980 Before Perl 5.28 this feature produced a warning in the
981 C<"experimental::bitwise"> category.
983 =head2 Bitwise Or and Exclusive Or
984 X<operator, bitwise, or> X<bitwise or> X<|> X<operator, bitwise, xor>
987 Binary C<"|"> returns its operands ORed together bit by bit.
989 Binary C<"^"> returns its operands XORed together bit by bit.
991 Although no warning is currently raised, the results are not well
992 defined when these operations are performed on operands that aren't either
993 numbers (see L</Integer Arithmetic>) nor bitstrings (see L</Bitwise String
996 Note that C<"|"> and C<"^"> have lower priority than relational operators, so
997 for example the parentheses are essential in a test like
999 print "false\n" if (8 | 2) != 10;
1001 If the "bitwise" feature is enabled via S<C<use feature 'bitwise'>> or
1002 C<use v5.28>, then this operator always treats its operands as numbers.
1003 Before Perl 5.28. this feature produced a warning in the
1004 C<"experimental::bitwise"> category.
1006 =head2 C-style Logical And
1007 X<&&> X<logical and> X<operator, logical, and>
1009 Binary C<"&&"> performs a short-circuit logical AND operation. That is,
1010 if the left operand is false, the right operand is not even evaluated.
1011 Scalar or list context propagates down to the right operand if it
1014 =head2 C-style Logical Or
1015 X<||> X<operator, logical, or>
1017 Binary C<"||"> performs a short-circuit logical OR operation. That is,
1018 if the left operand is true, the right operand is not even evaluated.
1019 Scalar or list context propagates down to the right operand if it
1022 =head2 Logical Defined-Or
1023 X<//> X<operator, logical, defined-or>
1025 Although it has no direct equivalent in C, Perl's C<//> operator is related
1026 to its C-style "or". In fact, it's exactly the same as C<||>, except that it
1027 tests the left hand side's definedness instead of its truth. Thus,
1028 S<C<< EXPR1 // EXPR2 >>> returns the value of C<< EXPR1 >> if it's defined,
1029 otherwise, the value of C<< EXPR2 >> is returned.
1030 (C<< EXPR1 >> is evaluated in scalar context, C<< EXPR2 >>
1031 in the context of C<< // >> itself). Usually,
1032 this is the same result as S<C<< defined(EXPR1) ? EXPR1 : EXPR2 >>> (except that
1033 the ternary-operator form can be used as a lvalue, while S<C<< EXPR1 // EXPR2 >>>
1034 cannot). This is very useful for
1035 providing default values for variables. If you actually want to test if
1036 at least one of C<$x> and C<$y> is defined, use S<C<defined($x // $y)>>.
1038 The C<||>, C<//> and C<&&> operators return the last value evaluated
1039 (unlike C's C<||> and C<&&>, which return 0 or 1). Thus, a reasonably
1040 portable way to find out the home directory might be:
1044 // (getpwuid($<))[7]
1045 // die "You're homeless!\n";
1047 In particular, this means that you shouldn't use this
1048 for selecting between two aggregates for assignment:
1050 @a = @b || @c; # This doesn't do the right thing
1051 @a = scalar(@b) || @c; # because it really means this.
1052 @a = @b ? @b : @c; # This works fine, though.
1054 As alternatives to C<&&> and C<||> when used for
1055 control flow, Perl provides the C<and> and C<or> operators (see below).
1056 The short-circuit behavior is identical. The precedence of C<"and">
1057 and C<"or"> is much lower, however, so that you can safely use them after a
1058 list operator without the need for parentheses:
1060 unlink "alpha", "beta", "gamma"
1061 or gripe(), next LINE;
1063 With the C-style operators that would have been written like this:
1065 unlink("alpha", "beta", "gamma")
1066 || (gripe(), next LINE);
1068 It would be even more readable to write that this way:
1070 unless(unlink("alpha", "beta", "gamma")) {
1075 Using C<"or"> for assignment is unlikely to do what you want; see below.
1077 =head2 Range Operators
1078 X<operator, range> X<range> X<..> X<...>
1080 Binary C<".."> is the range operator, which is really two different
1081 operators depending on the context. In list context, it returns a
1082 list of values counting (up by ones) from the left value to the right
1083 value. If the left value is greater than the right value then it
1084 returns the empty list. The range operator is useful for writing
1085 S<C<foreach (1..10)>> loops and for doing slice operations on arrays. In
1086 the current implementation, no temporary array is created when the
1087 range operator is used as the expression in C<foreach> loops, but older
1088 versions of Perl might burn a lot of memory when you write something
1091 for (1 .. 1_000_000) {
1095 The range operator also works on strings, using the magical
1096 auto-increment, see below.
1098 In scalar context, C<".."> returns a boolean value. The operator is
1099 bistable, like a flip-flop, and emulates the line-range (comma)
1100 operator of B<sed>, B<awk>, and various editors. Each C<".."> operator
1101 maintains its own boolean state, even across calls to a subroutine
1102 that contains it. It is false as long as its left operand is false.
1103 Once the left operand is true, the range operator stays true until the
1104 right operand is true, I<AFTER> which the range operator becomes false
1105 again. It doesn't become false till the next time the range operator
1106 is evaluated. It can test the right operand and become false on the
1107 same evaluation it became true (as in B<awk>), but it still returns
1108 true once. If you don't want it to test the right operand until the
1109 next evaluation, as in B<sed>, just use three dots (C<"...">) instead of
1110 two. In all other regards, C<"..."> behaves just like C<".."> does.
1112 The right operand is not evaluated while the operator is in the
1113 "false" state, and the left operand is not evaluated while the
1114 operator is in the "true" state. The precedence is a little lower
1115 than || and &&. The value returned is either the empty string for
1116 false, or a sequence number (beginning with 1) for true. The sequence
1117 number is reset for each range encountered. The final sequence number
1118 in a range has the string C<"E0"> appended to it, which doesn't affect
1119 its numeric value, but gives you something to search for if you want
1120 to exclude the endpoint. You can exclude the beginning point by
1121 waiting for the sequence number to be greater than 1.
1123 If either operand of scalar C<".."> is a constant expression,
1124 that operand is considered true if it is equal (C<==>) to the current
1125 input line number (the C<$.> variable).
1127 To be pedantic, the comparison is actually S<C<int(EXPR) == int(EXPR)>>,
1128 but that is only an issue if you use a floating point expression; when
1129 implicitly using C<$.> as described in the previous paragraph, the
1130 comparison is S<C<int(EXPR) == int($.)>> which is only an issue when C<$.>
1131 is set to a floating point value and you are not reading from a file.
1132 Furthermore, S<C<"span" .. "spat">> or S<C<2.18 .. 3.14>> will not do what
1133 you want in scalar context because each of the operands are evaluated
1134 using their integer representation.
1138 As a scalar operator:
1140 if (101 .. 200) { print; } # print 2nd hundred lines, short for
1141 # if ($. == 101 .. $. == 200) { print; }
1143 next LINE if (1 .. /^$/); # skip header lines, short for
1144 # next LINE if ($. == 1 .. /^$/);
1145 # (typically in a loop labeled LINE)
1147 s/^/> / if (/^$/ .. eof()); # quote body
1149 # parse mail messages
1151 $in_header = 1 .. /^$/;
1152 $in_body = /^$/ .. eof;
1159 close ARGV if eof; # reset $. each file
1162 Here's a simple example to illustrate the difference between
1163 the two range operators:
1176 This program will print only the line containing "Bar". If
1177 the range operator is changed to C<...>, it will also print the
1180 And now some examples as a list operator:
1182 for (101 .. 200) { print } # print $_ 100 times
1183 @foo = @foo[0 .. $#foo]; # an expensive no-op
1184 @foo = @foo[$#foo-4 .. $#foo]; # slice last 5 items
1186 Because each operand is evaluated in integer form, S<C<2.18 .. 3.14>> will
1187 return two elements in list context.
1189 @list = (2.18 .. 3.14); # same as @list = (2 .. 3);
1191 The range operator in list context can make use of the magical
1192 auto-increment algorithm if both operands are strings, subject to the
1199 With one exception (below), if both strings look like numbers to Perl,
1200 the magic increment will not be applied, and the strings will be treated
1201 as numbers (more specifically, integers) instead.
1203 For example, C<"-2".."2"> is the same as C<-2..2>, and
1204 C<"2.18".."3.14"> produces C<2, 3>.
1208 The exception to the above rule is when the left-hand string begins with
1209 C<0> and is longer than one character, in this case the magic increment
1210 I<will> be applied, even though strings like C<"01"> would normally look
1211 like a number to Perl.
1213 For example, C<"01".."04"> produces C<"01", "02", "03", "04">, and
1214 C<"00".."-1"> produces C<"00"> through C<"99"> - this may seem
1215 surprising, but see the following rules for why it works this way.
1216 To get dates with leading zeros, you can say:
1218 @z2 = ("01" .. "31");
1221 If you want to force strings to be interpreted as numbers, you could say
1223 @numbers = ( 0+$first .. 0+$last );
1225 B<Note:> In Perl versions 5.30 and below, I<any> string on the left-hand
1226 side beginning with C<"0">, including the string C<"0"> itself, would
1227 cause the magic string increment behavior. This means that on these Perl
1228 versions, C<"0".."-1"> would produce C<"0"> through C<"99">, which was
1229 inconsistent with C<0..-1>, which produces the empty list. This also means
1230 that C<"0".."9"> now produces a list of integers instead of a list of
1235 If the initial value specified isn't part of a magical increment
1236 sequence (that is, a non-empty string matching C</^[a-zA-Z]*[0-9]*\z/>),
1237 only the initial value will be returned.
1239 For example, C<"ax".."az"> produces C<"ax", "ay", "az">, but
1240 C<"*x".."az"> produces only C<"*x">.
1244 For other initial values that are strings that do follow the rules of the
1245 magical increment, the corresponding sequence will be returned.
1247 For example, you can say
1249 @alphabet = ("A" .. "Z");
1251 to get all normal letters of the English alphabet, or
1253 $hexdigit = (0 .. 9, "a" .. "f")[$num & 15];
1255 to get a hexadecimal digit.
1259 If the final value specified is not in the sequence that the magical
1260 increment would produce, the sequence goes until the next value would
1261 be longer than the final value specified. If the length of the final
1262 string is shorter than the first, the empty list is returned.
1264 For example, C<"a".."--"> is the same as C<"a".."zz">, C<"0".."xx">
1265 produces C<"0"> through C<"99">, and C<"aaa".."--"> returns the empty
1270 As of Perl 5.26, the list-context range operator on strings works as expected
1271 in the scope of L<< S<C<"use feature 'unicode_strings">>|feature/The
1272 'unicode_strings' feature >>. In previous versions, and outside the scope of
1273 that feature, it exhibits L<perlunicode/The "Unicode Bug">: its behavior
1274 depends on the internal encoding of the range endpoint.
1276 Because the magical increment only works on non-empty strings matching
1277 C</^[a-zA-Z]*[0-9]*\z/>, the following will only return an alpha:
1279 use charnames "greek";
1280 my @greek_small = ("\N{alpha}" .. "\N{omega}");
1282 To get the 25 traditional lowercase Greek letters, including both sigmas,
1283 you could use this instead:
1285 use charnames "greek";
1286 my @greek_small = map { chr } ( ord("\N{alpha}")
1291 However, because there are I<many> other lowercase Greek characters than
1292 just those, to match lowercase Greek characters in a regular expression,
1293 you could use the pattern C</(?:(?=\p{Greek})\p{Lower})+/> (or the
1294 L<experimental feature|perlrecharclass/Extended Bracketed Character
1295 Classes> C<S</(?[ \p{Greek} & \p{Lower} ])+/>>).
1297 =head2 Conditional Operator
1298 X<operator, conditional> X<operator, ternary> X<ternary> X<?:>
1300 Ternary C<"?:"> is the conditional operator, just as in C. It works much
1301 like an if-then-else. If the argument before the C<?> is true, the
1302 argument before the C<:> is returned, otherwise the argument after the
1303 C<:> is returned. For example:
1305 printf "I have %d dog%s.\n", $n,
1306 ($n == 1) ? "" : "s";
1308 Scalar or list context propagates downward into the 2nd
1309 or 3rd argument, whichever is selected.
1311 $x = $ok ? $y : $z; # get a scalar
1312 @x = $ok ? @y : @z; # get an array
1313 $x = $ok ? @y : @z; # oops, that's just a count!
1315 The operator may be assigned to if both the 2nd and 3rd arguments are
1316 legal lvalues (meaning that you can assign to them):
1318 ($x_or_y ? $x : $y) = $z;
1320 Because this operator produces an assignable result, using assignments
1321 without parentheses will get you in trouble. For example, this:
1323 $x % 2 ? $x += 10 : $x += 2
1327 (($x % 2) ? ($x += 10) : $x) += 2
1331 ($x % 2) ? ($x += 10) : ($x += 2)
1333 That should probably be written more simply as:
1335 $x += ($x % 2) ? 10 : 2;
1337 =head2 Assignment Operators
1338 X<assignment> X<operator, assignment> X<=> X<**=> X<+=> X<*=> X<&=>
1339 X<<< <<= >>> X<&&=> X<-=> X</=> X<|=> X<<< >>= >>> X<||=> X<//=> X<.=>
1340 X<%=> X<^=> X<x=> X<&.=> X<|.=> X<^.=>
1342 C<"="> is the ordinary assignment operator.
1344 Assignment operators work as in C. That is,
1352 although without duplicating any side effects that dereferencing the lvalue
1353 might trigger, such as from C<tie()>. Other assignment operators work similarly.
1354 The following are recognized:
1356 **= += *= &= &.= <<= &&=
1357 -= /= |= |.= >>= ||=
1361 Although these are grouped by family, they all have the precedence
1362 of assignment. These combined assignment operators can only operate on
1363 scalars, whereas the ordinary assignment operator can assign to arrays,
1364 hashes, lists and even references. (See L<"Context"|perldata/Context>
1365 and L<perldata/List value constructors>, and L<perlref/Assigning to
1368 Unlike in C, the scalar assignment operator produces a valid lvalue.
1369 Modifying an assignment is equivalent to doing the assignment and
1370 then modifying the variable that was assigned to. This is useful
1371 for modifying a copy of something, like this:
1373 ($tmp = $global) =~ tr/13579/24680/;
1375 Although as of 5.14, that can be also be accomplished this way:
1378 $tmp = ($global =~ tr/13579/24680/r);
1389 Similarly, a list assignment in list context produces the list of
1390 lvalues assigned to, and a list assignment in scalar context returns
1391 the number of elements produced by the expression on the right hand
1392 side of the assignment.
1394 The three dotted bitwise assignment operators (C<&.=> C<|.=> C<^.=>) are new in
1395 Perl 5.22. See L</Bitwise String Operators>.
1397 =head2 Comma Operator
1398 X<comma> X<operator, comma> X<,>
1400 Binary C<","> is the comma operator. In scalar context it evaluates
1401 its left argument, throws that value away, then evaluates its right
1402 argument and returns that value. This is just like C's comma operator.
1404 In list context, it's just the list argument separator, and inserts
1405 both its arguments into the list. These arguments are also evaluated
1408 The C<< => >> operator (sometimes pronounced "fat comma") is a synonym
1409 for the comma except that it causes a
1410 word on its left to be interpreted as a string if it begins with a letter
1411 or underscore and is composed only of letters, digits and underscores.
1412 This includes operands that might otherwise be interpreted as operators,
1413 constants, single number v-strings or function calls. If in doubt about
1414 this behavior, the left operand can be quoted explicitly.
1416 Otherwise, the C<< => >> operator behaves exactly as the comma operator
1417 or list argument separator, according to context.
1421 use constant FOO => "something";
1423 my %h = ( FOO => 23 );
1427 my %h = ("FOO", 23);
1431 my %h = ("something", 23);
1433 The C<< => >> operator is helpful in documenting the correspondence
1434 between keys and values in hashes, and other paired elements in lists.
1436 %hash = ( $key => $value );
1437 login( $username => $password );
1439 The special quoting behavior ignores precedence, and hence may apply to
1440 I<part> of the left operand:
1442 print time.shift => "bbb";
1444 That example prints something like C<"1314363215shiftbbb">, because the
1445 C<< => >> implicitly quotes the C<shift> immediately on its left, ignoring
1446 the fact that C<time.shift> is the entire left operand.
1448 =head2 List Operators (Rightward)
1449 X<operator, list, rightward> X<list operator>
1451 On the right side of a list operator, the comma has very low precedence,
1452 such that it controls all comma-separated expressions found there.
1453 The only operators with lower precedence are the logical operators
1454 C<"and">, C<"or">, and C<"not">, which may be used to evaluate calls to list
1455 operators without the need for parentheses:
1457 open HANDLE, "< :encoding(UTF-8)", "filename"
1458 or die "Can't open: $!\n";
1460 However, some people find that code harder to read than writing
1461 it with parentheses:
1463 open(HANDLE, "< :encoding(UTF-8)", "filename")
1464 or die "Can't open: $!\n";
1466 in which case you might as well just use the more customary C<"||"> operator:
1468 open(HANDLE, "< :encoding(UTF-8)", "filename")
1469 || die "Can't open: $!\n";
1471 See also discussion of list operators in L</Terms and List Operators (Leftward)>.
1474 X<operator, logical, not> X<not>
1476 Unary C<"not"> returns the logical negation of the expression to its right.
1477 It's the equivalent of C<"!"> except for the very low precedence.
1480 X<operator, logical, and> X<and>
1482 Binary C<"and"> returns the logical conjunction of the two surrounding
1483 expressions. It's equivalent to C<&&> except for the very low
1484 precedence. This means that it short-circuits: the right
1485 expression is evaluated only if the left expression is true.
1487 =head2 Logical or and Exclusive Or
1488 X<operator, logical, or> X<operator, logical, xor>
1489 X<operator, logical, exclusive or>
1492 Binary C<"or"> returns the logical disjunction of the two surrounding
1493 expressions. It's equivalent to C<||> except for the very low precedence.
1494 This makes it useful for control flow:
1496 print FH $data or die "Can't write to FH: $!";
1498 This means that it short-circuits: the right expression is evaluated
1499 only if the left expression is false. Due to its precedence, you must
1500 be careful to avoid using it as replacement for the C<||> operator.
1501 It usually works out better for flow control than in assignments:
1503 $x = $y or $z; # bug: this is wrong
1504 ($x = $y) or $z; # really means this
1505 $x = $y || $z; # better written this way
1507 However, when it's a list-context assignment and you're trying to use
1508 C<||> for control flow, you probably need C<"or"> so that the assignment
1509 takes higher precedence.
1511 @info = stat($file) || die; # oops, scalar sense of stat!
1512 @info = stat($file) or die; # better, now @info gets its due
1514 Then again, you could always use parentheses.
1516 Binary C<"xor"> returns the exclusive-OR of the two surrounding expressions.
1517 It cannot short-circuit (of course).
1519 There is no low precedence operator for defined-OR.
1521 =head2 C Operators Missing From Perl
1522 X<operator, missing from perl> X<&> X<*>
1523 X<typecasting> X<(TYPE)>
1525 Here is what C has that Perl doesn't:
1531 Address-of operator. (But see the C<"\"> operator for taking a reference.)
1535 Dereference-address operator. (Perl's prefix dereferencing
1536 operators are typed: C<$>, C<@>, C<%>, and C<&>.)
1540 Type-casting operator.
1544 =head2 Quote and Quote-like Operators
1545 X<operator, quote> X<operator, quote-like> X<q> X<qq> X<qx> X<qw> X<m>
1546 X<qr> X<s> X<tr> X<'> X<''> X<"> X<""> X<//> X<`> X<``> X<<< << >>>
1547 X<escape sequence> X<escape>
1549 While we usually think of quotes as literal values, in Perl they
1550 function as operators, providing various kinds of interpolating and
1551 pattern matching capabilities. Perl provides customary quote characters
1552 for these behaviors, but also provides a way for you to choose your
1553 quote character for any of them. In the following table, a C<{}> represents
1554 any pair of delimiters you choose.
1556 Customary Generic Meaning Interpolates
1559 `` qx{} Command yes*
1561 // m{} Pattern match yes*
1563 s{}{} Substitution yes*
1564 tr{}{} Transliteration no (but see below)
1565 y{}{} Transliteration no (but see below)
1568 * unless the delimiter is ''.
1570 Non-bracketing delimiters use the same character fore and aft, but the four
1571 sorts of ASCII brackets (round, angle, square, curly) all nest, which means
1580 Note, however, that this does not always work for quoting Perl code:
1582 $s = q{ if($x eq "}") ... }; # WRONG
1584 is a syntax error. The C<L<Text::Balanced>> module (standard as of v5.8,
1585 and from CPAN before then) is able to do this properly.
1587 There can (and in some cases, must) be whitespace between the operator
1589 characters, except when C<#> is being used as the quoting character.
1590 C<q#foo#> is parsed as the string C<foo>, while S<C<q #foo#>> is the
1591 operator C<q> followed by a comment. Its argument will be taken
1592 from the next line. This allows you to write:
1594 s {foo} # Replace foo
1597 The cases where whitespace must be used are when the quoting character
1598 is a word character (meaning it matches C</\w/>):
1600 q XfooX # Works: means the string 'foo'
1603 The following escape sequences are available in constructs that interpolate,
1604 and in transliterations whose delimiters aren't single quotes (C<"'">).
1605 In all the ones with braces, any number of blanks and/or tabs adjoining
1606 and within the braces are allowed (and ignored).
1607 X<\t> X<\n> X<\r> X<\f> X<\b> X<\a> X<\e> X<\x> X<\0> X<\c> X<\N> X<\N{}>
1610 Sequence Note Description
1616 \a alarm (bell) (BEL)
1618 \x{263A} [1,8] hex char (example shown: SMILEY)
1619 \x{ 263A } Same, but shows optional blanks inside and
1620 adjoining the braces
1621 \x1b [2,8] restricted range hex char (example: ESC)
1622 \N{name} [3] named Unicode character or character sequence
1623 \N{U+263D} [4,8] Unicode character (example: FIRST QUARTER MOON)
1624 \c[ [5] control char (example: chr(27))
1625 \o{23072} [6,8] octal char (example: SMILEY)
1626 \033 [7,8] restricted range octal char (example: ESC)
1628 Note that any escape sequence using braces inside interpolated
1629 constructs may have optional blanks (tab or space characters) adjoining
1630 with and inside of the braces, as illustrated above by the second
1631 S<C<\x{ }>> example.
1637 The result is the character specified by the hexadecimal number between
1638 the braces. See L</[8]> below for details on which character.
1640 Blanks (tab or space characters) may separate the number from either or
1643 Otherwise, only hexadecimal digits are valid between the braces. If an
1644 invalid character is encountered, a warning will be issued and the
1645 invalid character and all subsequent characters (valid or invalid)
1646 within the braces will be discarded.
1648 If there are no valid digits between the braces, the generated character is
1649 the NULL character (C<\x{00}>). However, an explicit empty brace (C<\x{}>)
1650 will not cause a warning (currently).
1654 The result is the character specified by the hexadecimal number in the range
1655 0x00 to 0xFF. See L</[8]> below for details on which character.
1657 Only hexadecimal digits are valid following C<\x>. When C<\x> is followed
1658 by fewer than two valid digits, any valid digits will be zero-padded. This
1659 means that C<\x7> will be interpreted as C<\x07>, and a lone C<"\x"> will be
1660 interpreted as C<\x00>. Except at the end of a string, having fewer than
1661 two valid digits will result in a warning. Note that although the warning
1662 says the illegal character is ignored, it is only ignored as part of the
1663 escape and will still be used as the subsequent character in the string.
1666 Original Result Warns?
1674 The result is the Unicode character or character sequence given by I<name>.
1679 S<C<\N{U+I<hexadecimal number>}>> means the Unicode character whose Unicode code
1680 point is I<hexadecimal number>.
1684 The character following C<\c> is mapped to some other character as shown in the
1697 # See below for chr(28)
1701 \c? chr(127) # (on ASCII platforms; see below for link to
1702 # EBCDIC discussion)
1704 In other words, it's the character whose code point has had 64 xor'd with
1705 its uppercase. C<\c?> is DELETE on ASCII platforms because
1706 S<C<ord("?") ^ 64>> is 127, and
1707 C<\c@> is NULL because the ord of C<"@"> is 64, so xor'ing 64 itself produces 0.
1709 Also, C<\c\I<X>> yields S<C< chr(28) . "I<X>">> for any I<X>, but cannot come at the
1710 end of a string, because the backslash would be parsed as escaping the end
1713 On ASCII platforms, the resulting characters from the list above are the
1714 complete set of ASCII controls. This isn't the case on EBCDIC platforms; see
1715 L<perlebcdic/OPERATOR DIFFERENCES> for a full discussion of the
1716 differences between these for ASCII versus EBCDIC platforms.
1718 Use of any other character following the C<"c"> besides those listed above is
1719 discouraged, and as of Perl v5.20, the only characters actually allowed
1720 are the printable ASCII ones, minus the left brace C<"{">. What happens
1721 for any of the allowed other characters is that the value is derived by
1722 xor'ing with the seventh bit, which is 64, and a warning raised if
1723 enabled. Using the non-allowed characters generates a fatal error.
1725 To get platform independent controls, you can use C<\N{...}>.
1729 The result is the character specified by the octal number between the braces.
1730 See L</[8]> below for details on which character.
1732 Blanks (tab or space characters) may separate the number from either or
1735 Otherwise, if a character that isn't an octal digit is encountered, a
1736 warning is raised, and the value is based on the octal digits before it,
1737 discarding it and all following characters up to the closing brace. It
1738 is a fatal error if there are no octal digits at all.
1742 The result is the character specified by the three-digit octal number in the
1743 range 000 to 777 (but best to not use above 077, see next paragraph). See
1744 L</[8]> below for details on which character.
1746 Some contexts allow 2 or even 1 digit, but any usage without exactly
1747 three digits, the first being a zero, may give unintended results. (For
1748 example, in a regular expression it may be confused with a backreference;
1749 see L<perlrebackslash/Octal escapes>.) Starting in Perl 5.14, you may
1750 use C<\o{}> instead, which avoids all these problems. Otherwise, it is best to
1751 use this construct only for ordinals C<\077> and below, remembering to pad to
1752 the left with zeros to make three digits. For larger ordinals, either use
1753 C<\o{}>, or convert to something else, such as to hex and use C<\N{U+}>
1754 (which is portable between platforms with different character sets) or
1759 Several constructs above specify a character by a number. That number
1760 gives the character's position in the character set encoding (indexed from 0).
1761 This is called synonymously its ordinal, code position, or code point. Perl
1762 works on platforms that have a native encoding currently of either ASCII/Latin1
1763 or EBCDIC, each of which allow specification of 256 characters. In general, if
1764 the number is 255 (0xFF, 0377) or below, Perl interprets this in the platform's
1765 native encoding. If the number is 256 (0x100, 0400) or above, Perl interprets
1766 it as a Unicode code point and the result is the corresponding Unicode
1767 character. For example C<\x{50}> and C<\o{120}> both are the number 80 in
1768 decimal, which is less than 256, so the number is interpreted in the native
1769 character set encoding. In ASCII the character in the 80th position (indexed
1770 from 0) is the letter C<"P">, and in EBCDIC it is the ampersand symbol C<"&">.
1771 C<\x{100}> and C<\o{400}> are both 256 in decimal, so the number is interpreted
1772 as a Unicode code point no matter what the native encoding is. The name of the
1773 character in the 256th position (indexed by 0) in Unicode is
1774 C<LATIN CAPITAL LETTER A WITH MACRON>.
1776 An exception to the above rule is that S<C<\N{U+I<hex number>}>> is
1777 always interpreted as a Unicode code point, so that C<\N{U+0050}> is C<"P"> even
1778 on EBCDIC platforms.
1782 B<NOTE>: Unlike C and other languages, Perl has no C<\v> escape sequence for
1783 the vertical tab (VT, which is 11 in both ASCII and EBCDIC), but you may
1784 use C<\N{VT}>, C<\ck>, C<\N{U+0b}>, or C<\x0b>. (C<\v>
1785 does have meaning in regular expression patterns in Perl, see L<perlre>.)
1787 The following escape sequences are available in constructs that interpolate,
1788 but not in transliterations.
1789 X<\l> X<\u> X<\L> X<\U> X<\E> X<\Q> X<\F>
1791 \l lowercase next character only
1792 \u titlecase (not uppercase!) next character only
1793 \L lowercase all characters till \E or end of string
1794 \U uppercase all characters till \E or end of string
1795 \F foldcase all characters till \E or end of string
1796 \Q quote (disable) pattern metacharacters till \E or
1798 \E end either case modification or quoted section
1799 (whichever was last seen)
1801 See L<perlfunc/quotemeta> for the exact definition of characters that
1802 are quoted by C<\Q>.
1804 C<\L>, C<\U>, C<\F>, and C<\Q> can stack, in which case you need one
1805 C<\E> for each. For example:
1807 say "This \Qquoting \ubusiness \Uhere isn't quite\E done yet,\E is it?";
1808 This quoting\ Business\ HERE\ ISN\'T\ QUITE\ done\ yet\, is it?
1810 If a S<C<use locale>> form that includes C<LC_CTYPE> is in effect (see
1811 L<perllocale>), the case map used by C<\l>, C<\L>, C<\u>, and C<\U> is
1812 taken from the current locale. If Unicode (for example, C<\N{}> or code
1813 points of 0x100 or beyond) is being used, the case map used by C<\l>,
1814 C<\L>, C<\u>, and C<\U> is as defined by Unicode. That means that
1815 case-mapping a single character can sometimes produce a sequence of
1817 Under S<C<use locale>>, C<\F> produces the same results as C<\L>
1818 for all locales but a UTF-8 one, where it instead uses the Unicode
1821 All systems use the virtual C<"\n"> to represent a line terminator,
1822 called a "newline". There is no such thing as an unvarying, physical
1823 newline character. It is only an illusion that the operating system,
1824 device drivers, C libraries, and Perl all conspire to preserve. Not all
1825 systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
1826 on the ancient Macs (pre-MacOS X) of yesteryear, these used to be reversed,
1827 and on systems without a line terminator,
1828 printing C<"\n"> might emit no actual data. In general, use C<"\n"> when
1829 you mean a "newline" for your system, but use the literal ASCII when you
1830 need an exact character. For example, most networking protocols expect
1831 and prefer a CR+LF (C<"\015\012"> or C<"\cM\cJ">) for line terminators,
1832 and although they often accept just C<"\012">, they seldom tolerate just
1833 C<"\015">. If you get in the habit of using C<"\n"> for networking,
1834 you may be burned some day.
1835 X<newline> X<line terminator> X<eol> X<end of line>
1838 For constructs that do interpolate, variables beginning with "C<$>"
1839 or "C<@>" are interpolated. Subscripted variables such as C<$a[3]> or
1840 C<< $href->{key}[0] >> are also interpolated, as are array and hash slices.
1841 But method calls such as C<< $obj->meth >> are not.
1843 Interpolating an array or slice interpolates the elements in order,
1844 separated by the value of C<$">, so is equivalent to interpolating
1845 S<C<join $", @array>>. "Punctuation" arrays such as C<@*> are usually
1846 interpolated only if the name is enclosed in braces C<@{*}>, but the
1847 arrays C<@_>, C<@+>, and C<@-> are interpolated even without braces.
1849 For double-quoted strings, the quoting from C<\Q> is applied after
1850 interpolation and escapes are processed.
1852 "abc\Qfoo\tbar$s\Exyz"
1856 "abc" . quotemeta("foo\tbar$s") . "xyz"
1858 For the pattern of regex operators (C<qr//>, C<m//> and C<s///>),
1859 the quoting from C<\Q> is applied after interpolation is processed,
1860 but before escapes are processed. This allows the pattern to match
1861 literally (except for C<$> and C<@>). For example, the following matches:
1865 Because C<$> or C<@> trigger interpolation, you'll need to use something
1866 like C</\Quser\E\@\Qhost/> to match them literally.
1868 Patterns are subject to an additional level of interpretation as a
1869 regular expression. This is done as a second pass, after variables are
1870 interpolated, so that regular expressions may be incorporated into the
1871 pattern from the variables. If this is not what you want, use C<\Q> to
1872 interpolate a variable literally.
1874 Apart from the behavior described above, Perl does not expand
1875 multiple levels of interpolation. In particular, contrary to the
1876 expectations of shell programmers, back-quotes do I<NOT> interpolate
1877 within double quotes, nor do single quotes impede evaluation of
1878 variables when used within double quotes.
1880 =head2 Regexp Quote-Like Operators
1883 Here are the quote-like operators that apply to pattern
1884 matching and related activities.
1888 =item C<qr/I<STRING>/msixpodualn>
1889 X<qr> X</i> X</m> X</o> X</s> X</x> X</p>
1891 This operator quotes (and possibly compiles) its I<STRING> as a regular
1892 expression. I<STRING> is interpolated the same way as I<PATTERN>
1893 in C<m/I<PATTERN>/>. If C<"'"> is used as the delimiter, no variable
1894 interpolation is done. Returns a Perl value which may be used instead of the
1895 corresponding C</I<STRING>/msixpodualn> expression. The returned value is a
1896 normalized version of the original pattern. It magically differs from
1897 a string containing the same characters: C<ref(qr/x/)> returns "Regexp";
1898 however, dereferencing it is not well defined (you currently get the
1899 normalized version of the original pattern, but this may change).
1904 $rex = qr/my.STRING/is;
1905 print $rex; # prints (?si-xm:my.STRING)
1912 The result may be used as a subpattern in a match:
1915 $string =~ /foo${re}bar/; # can be interpolated in other
1917 $string =~ $re; # or used standalone
1918 $string =~ /$re/; # or this way
1920 Since Perl may compile the pattern at the moment of execution of the C<qr()>
1921 operator, using C<qr()> may have speed advantages in some situations,
1922 notably if the result of C<qr()> is used standalone:
1925 my $patterns = shift;
1926 my @compiled = map qr/$_/i, @$patterns;
1929 foreach my $pat (@compiled) {
1930 $success = 1, last if /$pat/;
1936 Precompilation of the pattern into an internal representation at
1937 the moment of C<qr()> avoids the need to recompile the pattern every
1938 time a match C</$pat/> is attempted. (Perl has many other internal
1939 optimizations, but none would be triggered in the above example if
1940 we did not use C<qr()> operator.)
1942 Options (specified by the following modifiers) are:
1944 m Treat string as multiple lines.
1945 s Treat string as single line. (Make . match a newline)
1946 i Do case-insensitive pattern matching.
1947 x Use extended regular expressions; specifying two
1948 x's means \t and the SPACE character are ignored within
1949 square-bracketed character classes
1950 p When matching preserve a copy of the matched string so
1951 that ${^PREMATCH}, ${^MATCH}, ${^POSTMATCH} will be
1952 defined (ignored starting in v5.20) as these are always
1953 defined starting in that release
1954 o Compile pattern only once.
1955 a ASCII-restrict: Use ASCII for \d, \s, \w and [[:posix:]]
1956 character classes; specifying two a's adds the further
1957 restriction that no ASCII character will match a
1958 non-ASCII one under /i.
1959 l Use the current run-time locale's rules.
1960 u Use Unicode rules.
1961 d Use Unicode or native charset, as in 5.12 and earlier.
1962 n Non-capture mode. Don't let () fill in $1, $2, etc...
1964 If a precompiled pattern is embedded in a larger pattern then the effect
1965 of C<"msixpluadn"> will be propagated appropriately. The effect that the
1966 C</o> modifier has is not propagated, being restricted to those patterns
1967 explicitly using it.
1969 The C</a>, C</d>, C</l>, and C</u> modifiers (added in Perl 5.14)
1970 control the character set rules, but C</a> is the only one you are likely
1971 to want to specify explicitly; the other three are selected
1972 automatically by various pragmas.
1974 See L<perlre> for additional information on valid syntax for I<STRING>, and
1975 for a detailed look at the semantics of regular expressions. In
1976 particular, all modifiers except the largely obsolete C</o> are further
1977 explained in L<perlre/Modifiers>. C</o> is described in the next section.
1979 =item C<m/I<PATTERN>/msixpodualngc>
1980 X<m> X<operator, match>
1981 X<regexp, options> X<regexp> X<regex, options> X<regex>
1982 X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c>
1984 =item C</I<PATTERN>/msixpodualngc>
1986 Searches a string for a pattern match, and in scalar context returns
1987 true if it succeeds, false if it fails. If no string is specified
1988 via the C<=~> or C<!~> operator, the C<$_> string is searched. (The
1989 string specified with C<=~> need not be an lvalue--it may be the
1990 result of an expression evaluation, but remember the C<=~> binds
1991 rather tightly.) See also L<perlre>.
1993 Options are as described in C<qr//> above; in addition, the following match
1994 process modifiers are available:
1996 g Match globally, i.e., find all occurrences.
1997 c Do not reset search position on a failed match when /g is
2000 If C<"/"> is the delimiter then the initial C<m> is optional. With the C<m>
2001 you can use any pair of non-whitespace (ASCII) characters
2002 as delimiters. This is particularly useful for matching path names
2003 that contain C<"/">, to avoid LTS (leaning toothpick syndrome). If C<"?"> is
2004 the delimiter, then a match-only-once rule applies,
2005 described in C<m?I<PATTERN>?> below. If C<"'"> (single quote) is the delimiter,
2006 no variable interpolation is performed on the I<PATTERN>.
2007 When using a delimiter character valid in an identifier, whitespace is required
2010 I<PATTERN> may contain variables, which will be interpolated
2011 every time the pattern search is evaluated, except
2012 for when the delimiter is a single quote. (Note that C<$(>, C<$)>, and
2013 C<$|> are not interpolated because they look like end-of-string tests.)
2014 Perl will not recompile the pattern unless an interpolated
2015 variable that it contains changes. You can force Perl to skip the
2016 test and never recompile by adding a C</o> (which stands for "once")
2017 after the trailing delimiter.
2018 Once upon a time, Perl would recompile regular expressions
2019 unnecessarily, and this modifier was useful to tell it not to do so, in the
2020 interests of speed. But now, the only reasons to use C</o> are one of:
2026 The variables are thousands of characters long and you know that they
2027 don't change, and you need to wring out the last little bit of speed by
2028 having Perl skip testing for that. (There is a maintenance penalty for
2029 doing this, as mentioning C</o> constitutes a promise that you won't
2030 change the variables in the pattern. If you do change them, Perl won't
2035 you want the pattern to use the initial values of the variables
2036 regardless of whether they change or not. (But there are saner ways
2037 of accomplishing this than using C</o>.)
2041 If the pattern contains embedded code, such as
2044 $code = 'foo(?{ $x })';
2047 then perl will recompile each time, even though the pattern string hasn't
2048 changed, to ensure that the current value of C<$x> is seen each time.
2049 Use C</o> if you want to avoid this.
2053 The bottom line is that using C</o> is almost never a good idea.
2055 =item The empty pattern C<//>
2057 If the I<PATTERN> evaluates to the empty string, the last
2058 I<successfully> matched regular expression is used instead. In this
2059 case, only the C<g> and C<c> flags on the empty pattern are honored;
2060 the other flags are taken from the original pattern. If no match has
2061 previously succeeded, this will (silently) act instead as a genuine
2062 empty pattern (which will always match).
2064 Note that it's possible to confuse Perl into thinking C<//> (the empty
2065 regex) is really C<//> (the defined-or operator). Perl is usually pretty
2066 good about this, but some pathological cases might trigger this, such as
2067 C<$x///> (is that S<C<($x) / (//)>> or S<C<$x // />>?) and S<C<print $fh //>>
2068 (S<C<print $fh(//>> or S<C<print($fh //>>?). In all of these examples, Perl
2069 will assume you meant defined-or. If you meant the empty regex, just
2070 use parentheses or spaces to disambiguate, or even prefix the empty
2071 regex with an C<m> (so C<//> becomes C<m//>).
2073 =item Matching in list context
2075 If the C</g> option is not used, C<m//> in list context returns a
2076 list consisting of the subexpressions matched by the parentheses in the
2077 pattern, that is, (C<$1>, C<$2>, C<$3>...) (Note that here C<$1> etc. are
2078 also set). When there are no parentheses in the pattern, the return
2079 value is the list C<(1)> for success.
2080 With or without parentheses, an empty list is returned upon failure.
2084 open(TTY, "+</dev/tty")
2085 || die "can't access /dev/tty: $!";
2087 <TTY> =~ /^y/i && foo(); # do foo if desired
2089 if (/Version: *([0-9.]*)/) { $version = $1; }
2091 next if m#^/usr/spool/uucp#;
2096 print if /$arg/o; # compile only once (no longer needed!)
2099 if (($F1, $F2, $Etc) = ($foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
2101 This last example splits C<$foo> into the first two words and the
2102 remainder of the line, and assigns those three fields to C<$F1>, C<$F2>, and
2103 C<$Etc>. The conditional is true if any variables were assigned; that is,
2104 if the pattern matched.
2106 The C</g> modifier specifies global pattern matching--that is,
2107 matching as many times as possible within the string. How it behaves
2108 depends on the context. In list context, it returns a list of the
2109 substrings matched by any capturing parentheses in the regular
2110 expression. If there are no parentheses, it returns a list of all
2111 the matched strings, as if there were parentheses around the whole
2114 In scalar context, each execution of C<m//g> finds the next match,
2115 returning true if it matches, and false if there is no further match.
2116 The position after the last match can be read or set using the C<pos()>
2117 function; see L<perlfunc/pos>. A failed match normally resets the
2118 search position to the beginning of the string, but you can avoid that
2119 by adding the C</c> modifier (for example, C<m//gc>). Modifying the target
2120 string also resets the search position.
2122 =item C<\G I<assertion>>
2124 You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
2125 zero-width assertion that matches the exact position where the
2126 previous C<m//g>, if any, left off. Without the C</g> modifier, the
2127 C<\G> assertion still anchors at C<pos()> as it was at the start of
2128 the operation (see L<perlfunc/pos>), but the match is of course only
2129 attempted once. Using C<\G> without C</g> on a target string that has
2130 not previously had a C</g> match applied to it is the same as using
2131 the C<\A> assertion to match the beginning of the string. Note also
2132 that, currently, C<\G> is only properly supported when anchored at the
2133 very beginning of the pattern.
2138 ($one,$five,$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
2142 while ($paragraph = <>) {
2143 while ($paragraph =~ /\p{Ll}['")]*[.!?]+['")]*\s/g) {
2149 Here's another way to check for sentences in a paragraph:
2151 my $sentence_rx = qr{
2152 (?: (?<= ^ ) | (?<= \s ) ) # after start-of-string or
2154 \p{Lu} # capital letter
2155 .*? # a bunch of anything
2156 (?<= \S ) # that ends in non-
2158 (?<! \b [DMS]r ) # but isn't a common abbr.
2162 [.?!] # followed by a sentence
2164 (?= $ | \s ) # in front of end-of-string
2168 while (my $paragraph = <>) {
2169 say "NEW PARAGRAPH";
2171 while ($paragraph =~ /($sentence_rx)/g) {
2172 printf "\tgot sentence %d: <%s>\n", ++$count, $1;
2176 Here's how to use C<m//gc> with C<\G>:
2181 print $1 while /(o)/gc; print "', pos=", pos, "\n";
2183 print $1 if /\G(q)/gc; print "', pos=", pos, "\n";
2185 print $1 while /(p)/gc; print "', pos=", pos, "\n";
2187 print "Final: '$1', pos=",pos,"\n" if /\G(.)/;
2189 The last example should print:
2199 Notice that the final match matched C<q> instead of C<p>, which a match
2200 without the C<\G> anchor would have done. Also note that the final match
2201 did not update C<pos>. C<pos> is only updated on a C</g> match. If the
2202 final match did indeed match C<p>, it's a good bet that you're running an
2203 ancient (pre-5.6.0) version of Perl.
2205 A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
2206 combine several regexps like this to process a string part-by-part,
2207 doing different actions depending on which regexp matched. Each
2208 regexp tries to match where the previous one leaves off.
2211 $url = URI::URL->new( "http://example.com/" );
2212 die if $url eq "xXx";
2216 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
2217 print(" lowercase"), redo LOOP
2218 if /\G\p{Ll}+\b[,.;]?\s*/gc;
2219 print(" UPPERCASE"), redo LOOP
2220 if /\G\p{Lu}+\b[,.;]?\s*/gc;
2221 print(" Capitalized"), redo LOOP
2222 if /\G\p{Lu}\p{Ll}+\b[,.;]?\s*/gc;
2223 print(" MiXeD"), redo LOOP if /\G\pL+\b[,.;]?\s*/gc;
2224 print(" alphanumeric"), redo LOOP
2225 if /\G[\p{Alpha}\pN]+\b[,.;]?\s*/gc;
2226 print(" line-noise"), redo LOOP if /\G\W+/gc;
2227 print ". That's all!\n";
2230 Here is the output (split into several lines):
2232 line-noise lowercase line-noise UPPERCASE line-noise UPPERCASE
2233 line-noise lowercase line-noise lowercase line-noise lowercase
2234 lowercase line-noise lowercase lowercase line-noise lowercase
2235 lowercase line-noise MiXeD line-noise. That's all!
2237 =item C<m?I<PATTERN>?msixpodualngc>
2238 X<?> X<operator, match-once>
2240 This is just like the C<m/I<PATTERN>/> search, except that it matches
2241 only once between calls to the C<reset()> operator. This is a useful
2242 optimization when you want to see only the first occurrence of
2243 something in each file of a set of files, for instance. Only C<m??>
2244 patterns local to the current package are reset.
2248 # blank line between header and body
2251 reset if eof; # clear m?? status for next file
2254 Another example switched the first "latin1" encoding it finds
2255 to "utf8" in a pod file:
2257 s//utf8/ if m? ^ =encoding \h+ \K latin1 ?x;
2259 The match-once behavior is controlled by the match delimiter being
2260 C<?>; with any other delimiter this is the normal C<m//> operator.
2262 In the past, the leading C<m> in C<m?I<PATTERN>?> was optional, but omitting it
2263 would produce a deprecation warning. As of v5.22.0, omitting it produces a
2264 syntax error. If you encounter this construct in older code, you can just add
2267 =item C<s/I<PATTERN>/I<REPLACEMENT>/msixpodualngcer>
2268 X<s> X<substitute> X<substitution> X<replace> X<regexp, replace>
2269 X<regexp, substitute> X</m> X</s> X</i> X</x> X</p> X</o> X</g> X</c> X</e> X</r>
2271 Searches a string for a pattern, and if found, replaces that pattern
2272 with the replacement text and returns the number of substitutions
2273 made. Otherwise it returns false (a value that is both an empty string (C<"">)
2274 and numeric zero (C<0>) as described in L</Relational Operators>).
2276 If the C</r> (non-destructive) option is used then it runs the
2277 substitution on a copy of the string and instead of returning the
2278 number of substitutions, it returns the copy whether or not a
2279 substitution occurred. The original string is never changed when
2280 C</r> is used. The copy will always be a plain string, even if the
2281 input is an object or a tied variable.
2283 If no string is specified via the C<=~> or C<!~> operator, the C<$_>
2284 variable is searched and modified. Unless the C</r> option is used,
2285 the string specified must be a scalar variable, an array element, a
2286 hash element, or an assignment to one of those; that is, some sort of
2289 If the delimiter chosen is a single quote, no variable interpolation is
2290 done on either the I<PATTERN> or the I<REPLACEMENT>. Otherwise, if the
2291 I<PATTERN> contains a C<$> that looks like a variable rather than an
2292 end-of-string test, the variable will be interpolated into the pattern
2293 at run-time. If you want the pattern compiled only once the first time
2294 the variable is interpolated, use the C</o> option. If the pattern
2295 evaluates to the empty string, the last successfully executed regular
2296 expression is used instead. See L<perlre> for further explanation on these.
2298 Options are as with C<m//> with the addition of the following replacement
2301 e Evaluate the right side as an expression.
2302 ee Evaluate the right side as a string then eval the
2304 r Return substitution and leave the original string
2307 Any non-whitespace delimiter may replace the slashes. Add space after
2308 the C<s> when using a character allowed in identifiers. If single quotes
2309 are used, no interpretation is done on the replacement string (the C</e>
2310 modifier overrides this, however). Note that Perl treats backticks
2311 as normal delimiters; the replacement text is not evaluated as a command.
2312 If the I<PATTERN> is delimited by bracketing quotes, the I<REPLACEMENT> has
2313 its own pair of quotes, which may or may not be bracketing quotes, for example,
2314 C<s(foo)(bar)> or C<< s<foo>/bar/ >>. A C</e> will cause the
2315 replacement portion to be treated as a full-fledged Perl expression
2316 and evaluated right then and there. It is, however, syntax checked at
2317 compile-time. A second C<e> modifier will cause the replacement portion
2318 to be C<eval>ed before being run as a Perl expression.
2322 s/\bgreen\b/mauve/g; # don't change wintergreen
2324 $path =~ s|/usr/bin|/usr/local/bin|;
2326 s/Login: $foo/Login: $bar/; # run-time pattern
2328 ($foo = $bar) =~ s/this/that/; # copy first, then
2330 ($foo = "$bar") =~ s/this/that/; # convert to string,
2332 $foo = $bar =~ s/this/that/r; # Same as above using /r
2333 $foo = $bar =~ s/this/that/r
2334 =~ s/that/the other/r; # Chained substitutes
2336 @foo = map { s/this/that/r } @bar # /r is very useful in
2339 $count = ($paragraph =~ s/Mister\b/Mr./g); # get change-cnt
2342 s/\d+/$&*2/e; # yields 'abc246xyz'
2343 s/\d+/sprintf("%5d",$&)/e; # yields 'abc 246xyz'
2344 s/\w/$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
2346 s/%(.)/$percent{$1}/g; # change percent escapes; no /e
2347 s/%(.)/$percent{$1} || $&/ge; # expr now, so /e
2348 s/^=(\w+)/pod($1)/ge; # use function call
2351 $x = s/abc/def/r; # $x is 'def123xyz' and
2352 # $_ remains 'abc123xyz'.
2354 # expand variables in $_, but dynamics only, using
2355 # symbolic dereferencing
2358 # Add one to the value of any numbers in the string
2361 # Titlecase words in the last 30 characters only
2362 substr($str, -30) =~ s/\b(\p{Alpha}+)\b/\u\L$1/g;
2364 # This will expand any embedded scalar variable
2365 # (including lexicals) in $_ : First $1 is interpolated
2366 # to the variable name, and then evaluated
2369 # Delete (most) C comments.
2371 /\* # Match the opening delimiter.
2372 .*? # Match a minimal number of characters.
2373 \*/ # Match the closing delimiter.
2376 s/^\s*(.*?)\s*$/$1/; # trim whitespace in $_,
2379 for ($variable) { # trim whitespace in $variable,
2385 s/([^ ]*) *([^ ]*)/$2 $1/; # reverse 1st two fields
2387 $foo !~ s/A/a/g; # Lowercase all A's in $foo; return
2388 # 0 if any were found and changed;
2389 # otherwise return 1
2391 Note the use of C<$> instead of C<\> in the last example. Unlike
2392 B<sed>, we use the \<I<digit>> form only in the left hand side.
2393 Anywhere else it's $<I<digit>>.
2395 Occasionally, you can't use just a C</g> to get all the changes
2396 to occur that you might want. Here are two common cases:
2398 # put commas in the right places in an integer
2399 1 while s/(\d)(\d\d\d)(?!\d)/$1,$2/g;
2401 # expand tabs to 8-column spacing
2402 1 while s/\t+/' ' x (length($&)*8 - length($`)%8)/e;
2404 X</c>While C<s///> accepts the C</c> flag, it has no effect beyond
2405 producing a warning if warnings are enabled.
2409 =head2 Quote-Like Operators
2410 X<operator, quote-like>
2414 =item C<q/I<STRING>/>
2415 X<q> X<quote, single> X<'> X<''>
2417 =item C<'I<STRING>'>
2419 A single-quoted, literal string. A backslash represents a backslash
2420 unless followed by the delimiter or another backslash, in which case
2421 the delimiter or backslash is interpolated.
2423 $foo = q!I said, "You said, 'She said it.'"!;
2424 $bar = q('This is it.');
2425 $baz = '\n'; # a two-character string
2427 =item C<qq/I<STRING>/>
2428 X<qq> X<quote, double> X<"> X<"">
2430 =item C<"I<STRING>">
2432 A double-quoted, interpolated string.
2435 (*** The previous line contains the naughty word "$1".\n)
2436 if /\b(tcl|java|python)\b/i; # :-)
2437 $baz = "\n"; # a one-character string
2439 =item C<qx/I<STRING>/>
2440 X<qx> X<`> X<``> X<backtick>
2442 =item C<`I<STRING>`>
2444 A string which is (possibly) interpolated and then executed as a
2445 system command, via F</bin/sh> or its equivalent if required. Shell
2446 wildcards, pipes, and redirections will be honored. Similarly to
2447 C<system>, if the string contains no shell metacharacters then it will
2448 executed directly. The collected standard output of the command is
2449 returned; standard error is unaffected. In scalar context, it comes
2450 back as a single (potentially multi-line) string, or C<undef> if the
2451 shell (or command) could not be started. In list context, returns a
2452 list of lines (however you've defined lines with C<$/> or
2453 C<$INPUT_RECORD_SEPARATOR>), or an empty list if the shell (or command)
2454 could not be started.
2456 Because backticks do not affect standard error, use shell file descriptor
2457 syntax (assuming the shell supports this) if you care to address this.
2458 To capture a command's STDERR and STDOUT together:
2460 $output = `cmd 2>&1`;
2462 To capture a command's STDOUT but discard its STDERR:
2464 $output = `cmd 2>/dev/null`;
2466 To capture a command's STDERR but discard its STDOUT (ordering is
2469 $output = `cmd 2>&1 1>/dev/null`;
2471 To exchange a command's STDOUT and STDERR in order to capture the STDERR
2472 but leave its STDOUT to come out the old STDERR:
2474 $output = `cmd 3>&1 1>&2 2>&3 3>&-`;
2476 To read both a command's STDOUT and its STDERR separately, it's easiest
2477 to redirect them separately to files, and then read from those files
2478 when the program is done:
2480 system("program args 1>program.stdout 2>program.stderr");
2482 The STDIN filehandle used by the command is inherited from Perl's STDIN.
2485 open(SPLAT, "stuff") || die "can't open stuff: $!";
2486 open(STDIN, "<&SPLAT") || die "can't dupe SPLAT: $!";
2487 print STDOUT `sort`;
2489 will print the sorted contents of the file named F<"stuff">.
2491 Using single-quote as a delimiter protects the command from Perl's
2492 double-quote interpolation, passing it on to the shell instead:
2494 $perl_info = qx(ps $$); # that's Perl's $$
2495 $shell_info = qx'ps $$'; # that's the new shell's $$
2497 How that string gets evaluated is entirely subject to the command
2498 interpreter on your system. On most platforms, you will have to protect
2499 shell metacharacters if you want them treated literally. This is in
2500 practice difficult to do, as it's unclear how to escape which characters.
2501 See L<perlsec> for a clean and safe example of a manual C<fork()> and C<exec()>
2502 to emulate backticks safely.
2504 On some platforms (notably DOS-like ones), the shell may not be
2505 capable of dealing with multiline commands, so putting newlines in
2506 the string may not get you what you want. You may be able to evaluate
2507 multiple commands in a single line by separating them with the command
2508 separator character, if your shell supports that (for example, C<;> on
2509 many Unix shells and C<&> on the Windows NT C<cmd> shell).
2511 Perl will attempt to flush all files opened for
2512 output before starting the child process, but this may not be supported
2513 on some platforms (see L<perlport>). To be safe, you may need to set
2514 C<$|> (C<$AUTOFLUSH> in C<L<English>>) or call the C<autoflush()> method of
2515 C<L<IO::Handle>> on any open handles.
2517 Beware that some command shells may place restrictions on the length
2518 of the command line. You must ensure your strings don't exceed this
2519 limit after any necessary interpolations. See the platform-specific
2520 release notes for more details about your particular environment.
2522 Using this operator can lead to programs that are difficult to port,
2523 because the shell commands called vary between systems, and may in
2524 fact not be present at all. As one example, the C<type> command under
2525 the POSIX shell is very different from the C<type> command under DOS.
2526 That doesn't mean you should go out of your way to avoid backticks
2527 when they're the right way to get something done. Perl was made to be
2528 a glue language, and one of the things it glues together is commands.
2529 Just understand what you're getting yourself into.
2531 Like C<system>, backticks put the child process exit code in C<$?>.
2532 If you'd like to manually inspect failure, you can check all possible
2533 failure modes by inspecting C<$?> like this:
2536 print "failed to execute: $!\n";
2539 printf "child died with signal %d, %s coredump\n",
2540 ($? & 127), ($? & 128) ? 'with' : 'without';
2543 printf "child exited with value %d\n", $? >> 8;
2546 Use the L<open> pragma to control the I/O layers used when reading the
2547 output of the command, for example:
2549 use open IN => ":encoding(UTF-8)";
2550 my $x = `cmd-producing-utf-8`;
2552 C<qx//> can also be called like a function with L<perlfunc/readpipe>.
2554 See L</"I/O Operators"> for more discussion.
2556 =item C<qw/I<STRING>/>
2557 X<qw> X<quote, list> X<quote, words>
2559 Evaluates to a list of the words extracted out of I<STRING>, using embedded
2560 whitespace as the word delimiters. It can be understood as being roughly
2563 split(" ", q/STRING/);
2565 the differences being that it only splits on ASCII whitespace,
2566 generates a real list at compile time, and
2567 in scalar context it returns the last element in the list. So
2572 is semantically equivalent to the list:
2576 Some frequently seen examples:
2578 use POSIX qw( setlocale localeconv )
2579 @EXPORT = qw( foo bar baz );
2581 A common mistake is to try to separate the words with commas or to
2582 put comments into a multi-line C<qw>-string. For this reason, the
2583 S<C<use warnings>> pragma and the B<-w> switch (that is, the C<$^W> variable)
2584 produces warnings if the I<STRING> contains the C<","> or the C<"#"> character.
2586 =item C<tr/I<SEARCHLIST>/I<REPLACEMENTLIST>/cdsr>
2587 X<tr> X<y> X<transliterate> X</c> X</d> X</s>
2589 =item C<y/I<SEARCHLIST>/I<REPLACEMENTLIST>/cdsr>
2591 Transliterates all occurrences of the characters found (or not found
2592 if the C</c> modifier is specified) in the search list with the
2593 positionally corresponding character in the replacement list, possibly
2594 deleting some, depending on the modifiers specified. It returns the
2595 number of characters replaced or deleted. If no string is specified via
2596 the C<=~> or C<!~> operator, the C<$_> string is transliterated.
2598 For B<sed> devotees, C<y> is provided as a synonym for C<tr>.
2600 If the C</r> (non-destructive) option is present, a new copy of the string
2601 is made and its characters transliterated, and this copy is returned no
2602 matter whether it was modified or not: the original string is always
2603 left unchanged. The new copy is always a plain string, even if the input
2604 string is an object or a tied variable.
2606 Unless the C</r> option is used, the string specified with C<=~> must be a
2607 scalar variable, an array element, a hash element, or an assignment to one
2608 of those; in other words, an lvalue.
2610 The characters delimitting I<SEARCHLIST> and I<REPLACEMENTLIST>
2611 can be any printable character, not just forward slashes. If they
2612 are single quotes (C<tr'I<SEARCHLIST>'I<REPLACEMENTLIST>'>), the only
2613 interpolation is removal of C<\> from pairs of C<\\>, and hyphens are
2614 interpreted literally rather than specifying a character range.
2616 Otherwise, a character range may be specified with a hyphen, so
2617 C<tr/A-J/0-9/> does the same replacement as
2618 C<tr/ACEGIBDFHJ/0246813579/>.
2620 If the I<SEARCHLIST> is delimited by bracketing quotes, the
2621 I<REPLACEMENTLIST> must have its own pair of quotes, which may or may
2622 not be bracketing quotes; for example, C<tr[aeiouy][yuoiea]> or
2625 Characters may be literals, or (if the delimiters aren't single quotes)
2626 any of the escape sequences accepted in double-quoted strings. But
2627 there is never any variable interpolation, so C<"$"> and C<"@"> are
2628 always treated as literals. A hyphen at the beginning or end, or
2629 preceded by a backslash is also always considered a literal. Escape
2630 sequence details are in L<the table near the beginning of this
2631 section|/Quote and Quote-like Operators>.
2633 Note that C<tr> does B<not> do regular expression character classes such as
2634 C<\d> or C<\pL>. The C<tr> operator is not equivalent to the C<L<tr(1)>>
2635 utility. C<tr[a-z][A-Z]> will uppercase the 26 letters "a" through "z",
2636 but for case changing not confined to ASCII, use
2637 L<C<lc>|perlfunc/lc>, L<C<uc>|perlfunc/uc>,
2638 L<C<lcfirst>|perlfunc/lcfirst>, L<C<ucfirst>|perlfunc/ucfirst>
2639 (all documented in L<perlfunc>), or the
2640 L<substitution operator C<sE<sol>I<PATTERN>E<sol>I<REPLACEMENT>E<sol>>|/sE<sol>PATTERNE<sol>REPLACEMENTE<sol>msixpodualngcer>
2641 (with C<\U>, C<\u>, C<\L>, and C<\l> string-interpolation escapes in the
2642 I<REPLACEMENT> portion).
2644 Most ranges are unportable between character sets, but certain ones
2645 signal Perl to do special handling to make them portable. There are two
2646 classes of portable ranges. The first are any subsets of the ranges
2647 C<A-Z>, C<a-z>, and C<0-9>, when expressed as literal characters.
2651 capitalizes the letters C<"h">, C<"i">, C<"j">, and C<"k"> and nothing
2652 else, no matter what the platform's character set is. In contrast, all
2655 tr/\x68-\x6B/\x48-\x4B/
2659 do the same capitalizations as the previous example when run on ASCII
2660 platforms, but something completely different on EBCDIC ones.
2662 The second class of portable ranges is invoked when one or both of the
2663 range's end points are expressed as C<\N{...}>
2665 $string =~ tr/\N{U+20}-\N{U+7E}//d;
2667 removes from C<$string> all the platform's characters which are
2668 equivalent to any of Unicode U+0020, U+0021, ... U+007D, U+007E. This
2669 is a portable range, and has the same effect on every platform it is
2670 run on. In this example, these are the ASCII
2671 printable characters. So after this is run, C<$string> has only
2672 controls and characters which have no ASCII equivalents.
2674 But, even for portable ranges, it is not generally obvious what is
2675 included without having to look things up in the manual. A sound
2676 principle is to use only ranges that both begin from, and end at, either
2677 ASCII alphabetics of equal case (C<b-e>, C<B-E>), or digits (C<1-4>).
2678 Anything else is unclear (and unportable unless C<\N{...}> is used). If
2679 in doubt, spell out the character sets in full.
2683 c Complement the SEARCHLIST.
2684 d Delete found but unreplaced characters.
2685 r Return the modified string and leave the original string
2687 s Squash duplicate replaced characters.
2689 If the C</d> modifier is specified, any characters specified by
2690 I<SEARCHLIST> not found in I<REPLACEMENTLIST> are deleted. (Note that
2691 this is slightly more flexible than the behavior of some B<tr> programs,
2692 which delete anything they find in the I<SEARCHLIST>, period.)
2694 If the C</s> modifier is specified, sequences of characters, all in a
2695 row, that were transliterated to the same character are squashed down to
2696 a single instance of that character.
2699 $a =~ tr/ab/dd/s; # $a now is "dcd"
2701 If the C</d> modifier is used, the I<REPLACEMENTLIST> is always interpreted
2702 exactly as specified. Otherwise, if the I<REPLACEMENTLIST> is shorter
2703 than the I<SEARCHLIST>, the final character, if any, is replicated until
2704 it is long enough. There won't be a final character if and only if the
2705 I<REPLACEMENTLIST> is empty, in which case I<REPLACEMENTLIST> is
2706 copied from I<SEARCHLIST>. An empty I<REPLACEMENTLIST> is useful
2707 for counting characters in a class, or for squashing character sequences
2710 tr/abcd// tr/abcd/abcd/
2711 tr/abcd/AB/ tr/abcd/ABBB/
2712 tr/abcd//d s/[abcd]//g
2713 tr/abcd/AB/d (tr/ab/AB/ + s/[cd]//g) - but run together
2715 If the C</c> modifier is specified, the characters to be transliterated
2716 are the ones NOT in I<SEARCHLIST>, that is, it is complemented. If
2717 C</d> and/or C</s> are also specified, they apply to the complemented
2718 I<SEARCHLIST>. Recall, that if I<REPLACEMENTLIST> is empty (except
2719 under C</d>) a copy of I<SEARCHLIST> is used instead. That copy is made
2720 after complementing under C</c>. I<SEARCHLIST> is sorted by code point
2721 order after complementing, and any I<REPLACEMENTLIST> is applied to
2722 that sorted result. This means that under C</c>, the order of the
2723 characters specified in I<SEARCHLIST> is irrelevant. This can
2724 lead to different results on EBCDIC systems if I<REPLACEMENTLIST>
2725 contains more than one character, hence it is generally non-portable to
2726 use C</c> with such a I<REPLACEMENTLIST>.
2728 Another way of describing the operation is this:
2729 If C</c> is specified, the I<SEARCHLIST> is sorted by code point order,
2730 then complemented. If I<REPLACEMENTLIST> is empty and C</d> is not
2731 specified, I<REPLACEMENTLIST> is replaced by a copy of I<SEARCHLIST> (as
2732 modified under C</c>), and these potentially modified lists are used as
2733 the basis for what follows. Any character in the target string that
2734 isn't in I<SEARCHLIST> is passed through unchanged. Every other
2735 character in the target string is replaced by the character in
2736 I<REPLACEMENTLIST> that positionally corresponds to its mate in
2737 I<SEARCHLIST>, except that under C</s>, the 2nd and following characters
2738 are squeezed out in a sequence of characters in a row that all translate
2739 to the same character. If I<SEARCHLIST> is longer than
2740 I<REPLACEMENTLIST>, characters in the target string that match a
2741 character in I<SEARCHLIST> that doesn't have a correspondence in
2742 I<REPLACEMENTLIST> are either deleted from the target string if C</d> is
2743 specified; or replaced by the final character in I<REPLACEMENTLIST> if
2744 C</d> isn't specified.
2748 $ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case ASCII
2750 $cnt = tr/*/*/; # count the stars in $_
2751 $cnt = tr/*//; # same thing
2753 $cnt = $sky =~ tr/*/*/; # count the stars in $sky
2754 $cnt = $sky =~ tr/*//; # same thing
2756 $cnt = $sky =~ tr/*//c; # count all the non-stars in $sky
2757 $cnt = $sky =~ tr/*/*/c; # same, but transliterate each non-star
2758 # into a star, leaving the already-stars
2759 # alone. Afterwards, everything in $sky
2762 $cnt = tr/0-9//; # count the ASCII digits in $_
2764 tr/a-zA-Z//s; # bookkeeper -> bokeper
2765 tr/o/o/s; # bookkeeper -> bokkeeper
2766 tr/oe/oe/s; # bookkeeper -> bokkeper
2767 tr/oe//s; # bookkeeper -> bokkeper
2768 tr/oe/o/s; # bookkeeper -> bokkopor
2770 ($HOST = $host) =~ tr/a-z/A-Z/;
2771 $HOST = $host =~ tr/a-z/A-Z/r; # same thing
2773 $HOST = $host =~ tr/a-z/A-Z/r # chained with s///r
2776 tr/a-zA-Z/ /cs; # change non-alphas to single space
2778 @stripped = map tr/a-zA-Z/ /csr, @original;
2782 [\000-\177]; # wickedly delete 8th bit
2784 $foo !~ tr/A/a/ # transliterate all the A's in $foo to 'a',
2785 # return 0 if any were found and changed.
2786 # Otherwise return 1
2788 If multiple transliterations are given for a character, only the
2793 will transliterate any A to X.
2795 Because the transliteration table is built at compile time, neither
2796 the I<SEARCHLIST> nor the I<REPLACEMENTLIST> are subjected to double quote
2797 interpolation. That means that if you want to use variables, you
2798 must use an C<eval()>:
2800 eval "tr/$oldlist/$newlist/";
2803 eval "tr/$oldlist/$newlist/, 1" or die $@;
2805 =item C<< <<I<EOF> >>
2806 X<here-doc> X<heredoc> X<here-document> X<<< << >>>
2808 A line-oriented form of quoting is based on the shell "here-document"
2809 syntax. Following a C<< << >> you specify a string to terminate
2810 the quoted material, and all lines following the current line down to
2811 the terminating string are the value of the item.
2813 Prefixing the terminating string with a C<~> specifies that you
2814 want to use L</Indented Here-docs> (see below).
2816 The terminating string may be either an identifier (a word), or some
2817 quoted text. An unquoted identifier works like double quotes.
2818 There may not be a space between the C<< << >> and the identifier,
2819 unless the identifier is explicitly quoted. The terminating string
2820 must appear by itself (unquoted and with no surrounding whitespace)
2821 on the terminating line.
2823 If the terminating string is quoted, the type of quotes used determine
2824 the treatment of the text.
2830 Double quotes indicate that the text will be interpolated using exactly
2831 the same rules as normal double quoted strings.
2834 The price is $Price.
2837 print << "EOF"; # same as above
2838 The price is $Price.
2844 Single quotes indicate the text is to be treated literally with no
2845 interpolation of its content. This is similar to single quoted
2846 strings except that backslashes have no special meaning, with C<\\>
2847 being treated as two backslashes and not one as they would in every
2848 other quoting construct.
2850 Just as in the shell, a backslashed bareword following the C<<< << >>>
2851 means the same thing as a single-quoted string does:
2853 $cost = <<'VISTA'; # hasta la ...
2854 That'll be $10 please, ma'am.
2857 $cost = <<\VISTA; # Same thing!
2858 That'll be $10 please, ma'am.
2861 This is the only form of quoting in perl where there is no need
2862 to worry about escaping content, something that code generators
2863 can and do make good use of.
2867 The content of the here doc is treated just as it would be if the
2868 string were embedded in backticks. Thus the content is interpolated
2869 as though it were double quoted and then executed via the shell, with
2870 the results of the execution returned.
2872 print << `EOC`; # execute command and get results
2880 =item Indented Here-docs
2882 The here-doc modifier C<~> allows you to indent your here-docs to make
2883 the code more readable:
2895 ...with no leading whitespace.
2897 The line containing the delimiter that marks the end of the here-doc
2898 determines the indentation template for the whole thing. Compilation
2899 croaks if any non-empty line inside the here-doc does not begin with the
2900 precise indentation of the terminating line. (An empty line consists of
2901 the single character "\n".) For example, suppose the terminating line
2902 begins with a tab character followed by 4 space characters. Every
2903 non-empty line in the here-doc must begin with a tab followed by 4
2904 spaces. They are stripped from each line, and any leading white space
2905 remaining on a line serves as the indentation for that line. Currently,
2906 only the TAB and SPACE characters are treated as whitespace for this
2907 purpose. Tabs and spaces may be mixed, but are matched exactly; tabs
2908 remain tabs and are not expanded.
2910 Additional beginning whitespace (beyond what preceded the
2911 delimiter) will be preserved:
2914 This text is not indented
2915 This text is indented with two spaces
2916 This text is indented with two tabs
2919 Finally, the modifier may be used with all of the forms
2927 And whitespace may be used between the C<~> and quoted delimiters:
2929 <<~ 'EOF'; # ... "EOF", `EOF`
2933 It is possible to stack multiple here-docs in a row:
2935 print <<"foo", <<"bar"; # you can stack them
2941 myfunc(<< "THIS", 23, <<'THAT');
2948 Just don't forget that you have to put a semicolon on the end
2949 to finish the statement, as Perl doesn't know you're not going to
2957 If you want to remove the line terminator from your here-docs,
2960 chomp($string = <<'END');
2964 If you want your here-docs to be indented with the rest of the code,
2965 use the C<<< <<~FOO >>> construct described under L</Indented Here-docs>:
2967 $quote = <<~'FINIS';
2968 The Road goes ever on and on,
2969 down from the door where it began.
2972 If you use a here-doc within a delimited construct, such as in C<s///eg>,
2973 the quoted material must still come on the line following the
2974 C<<< <<FOO >>> marker, which means it may be inside the delimited
2982 It works this way as of Perl 5.18. Historically, it was inconsistent, and
2983 you would have to write
2990 outside of string evals.
2992 Additionally, quoting rules for the end-of-string identifier are
2993 unrelated to Perl's quoting rules. C<q()>, C<qq()>, and the like are not
2994 supported in place of C<''> and C<"">, and the only interpolation is for
2995 backslashing the quoting character:
2997 print << "abc\"def";
3001 Finally, quoted strings cannot span multiple lines. The general rule is
3002 that the identifier must be a string literal. Stick with that, and you
3007 =head2 Gory details of parsing quoted constructs
3008 X<quote, gory details>
3010 When presented with something that might have several different
3011 interpretations, Perl uses the B<DWIM> (that's "Do What I Mean")
3012 principle to pick the most probable interpretation. This strategy
3013 is so successful that Perl programmers often do not suspect the
3014 ambivalence of what they write. But from time to time, Perl's
3015 notions differ substantially from what the author honestly meant.
3017 This section hopes to clarify how Perl handles quoted constructs.
3018 Although the most common reason to learn this is to unravel labyrinthine
3019 regular expressions, because the initial steps of parsing are the
3020 same for all quoting operators, they are all discussed together.
3022 The most important Perl parsing rule is the first one discussed
3023 below: when processing a quoted construct, Perl first finds the end
3024 of that construct, then interprets its contents. If you understand
3025 this rule, you may skip the rest of this section on the first
3026 reading. The other rules are likely to contradict the user's
3027 expectations much less frequently than this first one.
3029 Some passes discussed below are performed concurrently, but because
3030 their results are the same, we consider them individually. For different
3031 quoting constructs, Perl performs different numbers of passes, from
3032 one to four, but these passes are always performed in the same order.
3036 =item Finding the end
3038 The first pass is finding the end of the quoted construct. This results
3039 in saving to a safe location a copy of the text (between the starting
3040 and ending delimiters), normalized as necessary to avoid needing to know
3041 what the original delimiters were.
3043 If the construct is a here-doc, the ending delimiter is a line
3044 that has a terminating string as the content. Therefore C<<<EOF> is
3045 terminated by C<EOF> immediately followed by C<"\n"> and starting
3046 from the first column of the terminating line.
3047 When searching for the terminating line of a here-doc, nothing
3048 is skipped. In other words, lines after the here-doc syntax
3049 are compared with the terminating string line by line.
3051 For the constructs except here-docs, single characters are used as starting
3052 and ending delimiters. If the starting delimiter is an opening punctuation
3053 (that is C<(>, C<[>, C<{>, or C<< < >>), the ending delimiter is the
3054 corresponding closing punctuation (that is C<)>, C<]>, C<}>, or C<< > >>).
3055 If the starting delimiter is an unpaired character like C</> or a closing
3056 punctuation, the ending delimiter is the same as the starting delimiter.
3057 Therefore a C</> terminates a C<qq//> construct, while a C<]> terminates
3058 both C<qq[]> and C<qq]]> constructs.
3060 When searching for single-character delimiters, escaped delimiters
3061 and C<\\> are skipped. For example, while searching for terminating C</>,
3062 combinations of C<\\> and C<\/> are skipped. If the delimiters are
3063 bracketing, nested pairs are also skipped. For example, while searching
3064 for a closing C<]> paired with the opening C<[>, combinations of C<\\>, C<\]>,
3065 and C<\[> are all skipped, and nested C<[> and C<]> are skipped as well.
3066 However, when backslashes are used as the delimiters (like C<qq\\> and
3067 C<tr\\\>), nothing is skipped.
3068 During the search for the end, backslashes that escape delimiters or
3069 other backslashes are removed (exactly speaking, they are not copied to the
3072 For constructs with three-part delimiters (C<s///>, C<y///>, and
3073 C<tr///>), the search is repeated once more.
3074 If the first delimiter is not an opening punctuation, the three delimiters must
3075 be the same, such as C<s!!!> and C<tr)))>,
3076 in which case the second delimiter
3077 terminates the left part and starts the right part at once.
3078 If the left part is delimited by bracketing punctuation (that is C<()>,
3079 C<[]>, C<{}>, or C<< <> >>), the right part needs another pair of
3080 delimiters such as C<s(){}> and C<tr[]//>. In these cases, whitespace
3081 and comments are allowed between the two parts, although the comment must follow
3082 at least one whitespace character; otherwise a character expected as the
3083 start of the comment may be regarded as the starting delimiter of the right part.
3085 During this search no attention is paid to the semantics of the construct.
3088 "$hash{"$foo/$bar"}"
3093 bar # NOT a comment, this slash / terminated m//!
3096 do not form legal quoted expressions. The quoted part ends on the
3097 first C<"> and C</>, and the rest happens to be a syntax error.
3098 Because the slash that terminated C<m//> was followed by a C<SPACE>,
3099 the example above is not C<m//x>, but rather C<m//> with no C</x>
3100 modifier. So the embedded C<#> is interpreted as a literal C<#>.
3102 Also no attention is paid to C<\c\> (multichar control char syntax) during
3103 this search. Thus the second C<\> in C<qq/\c\/> is interpreted as a part
3104 of C<\/>, and the following C</> is not recognized as a delimiter.
3105 Instead, use C<\034> or C<\x1c> at the end of quoted constructs.
3110 The next step is interpolation in the text obtained, which is now
3111 delimiter-independent. There are multiple cases.
3117 No interpolation is performed.
3118 Note that the combination C<\\> is left intact, since escaped delimiters
3119 are not available for here-docs.
3121 =item C<m''>, the pattern of C<s'''>
3123 No interpolation is performed at this stage.
3124 Any backslashed sequences including C<\\> are treated at the stage
3125 of L</"Parsing regular expressions">.
3127 =item C<''>, C<q//>, C<tr'''>, C<y'''>, the replacement of C<s'''>
3129 The only interpolation is removal of C<\> from pairs of C<\\>.
3130 Therefore C<"-"> in C<tr'''> and C<y'''> is treated literally
3131 as a hyphen and no character range is available.
3132 C<\1> in the replacement of C<s'''> does not work as C<$1>.
3134 =item C<tr///>, C<y///>
3136 No variable interpolation occurs. String modifying combinations for
3137 case and quoting such as C<\Q>, C<\U>, and C<\E> are not recognized.
3138 The other escape sequences such as C<\200> and C<\t> and backslashed
3139 characters such as C<\\> and C<\-> are converted to appropriate literals.
3140 The character C<"-"> is treated specially and therefore C<\-> is treated
3141 as a literal C<"-">.
3143 =item C<"">, C<``>, C<qq//>, C<qx//>, C<< <file*glob> >>, C<<<"EOF">
3145 C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\F> (possibly paired with C<\E>) are
3146 converted to corresponding Perl constructs. Thus, C<"$foo\Qbaz$bar">
3147 is converted to S<C<$foo . (quotemeta("baz" . $bar))>> internally.
3148 The other escape sequences such as C<\200> and C<\t> and backslashed
3149 characters such as C<\\> and C<\-> are replaced with appropriate
3152 Let it be stressed that I<whatever falls between C<\Q> and C<\E>>
3153 is interpolated in the usual way. Something like C<"\Q\\E"> has
3154 no C<\E> inside. Instead, it has C<\Q>, C<\\>, and C<E>, so the
3155 result is the same as for C<"\\\\E">. As a general rule, backslashes
3156 between C<\Q> and C<\E> may lead to counterintuitive results. So,
3157 C<"\Q\t\E"> is converted to C<quotemeta("\t")>, which is the same
3158 as C<"\\\t"> (since TAB is not alphanumeric). Note also that:
3163 may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
3165 Interpolated scalars and arrays are converted internally to the C<join> and
3166 C<"."> catenation operations. Thus, S<C<"$foo XXX '@arr'">> becomes:
3168 $foo . " XXX '" . (join $", @arr) . "'";
3170 All operations above are performed simultaneously, left to right.
3172 Because the result of S<C<"\Q I<STRING> \E">> has all metacharacters
3173 quoted, there is no way to insert a literal C<$> or C<@> inside a
3174 C<\Q\E> pair. If protected by C<\>, C<$> will be quoted to become
3175 C<"\\\$">; if not, it is interpreted as the start of an interpolated
3178 Note also that the interpolation code needs to make a decision on
3179 where the interpolated scalar ends. For instance, whether
3180 S<C<< "a $x -> {c}" >>> really means:
3182 "a " . $x . " -> {c}";
3188 Most of the time, the longest possible text that does not include
3189 spaces between components and which contains matching braces or
3190 brackets. because the outcome may be determined by voting based
3191 on heuristic estimators, the result is not strictly predictable.
3192 Fortunately, it's usually correct for ambiguous cases.
3194 =item The replacement of C<s///>
3196 Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\F> and interpolation
3197 happens as with C<qq//> constructs.
3199 It is at this step that C<\1> is begrudgingly converted to C<$1> in
3200 the replacement text of C<s///>, in order to correct the incorrigible
3201 I<sed> hackers who haven't picked up the saner idiom yet. A warning
3202 is emitted if the S<C<use warnings>> pragma or the B<-w> command-line flag
3203 (that is, the C<$^W> variable) was set.
3205 =item C<RE> in C<m?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
3207 Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, C<\F>, C<\E>,
3208 and interpolation happens (almost) as with C<qq//> constructs.
3210 Processing of C<\N{...}> is also done here, and compiled into an intermediate
3211 form for the regex compiler. (This is because, as mentioned below, the regex
3212 compilation may be done at execution time, and C<\N{...}> is a compile-time
3215 However any other combinations of C<\> followed by a character
3216 are not substituted but only skipped, in order to parse them
3217 as regular expressions at the following step.
3218 As C<\c> is skipped at this step, C<@> of C<\c@> in RE is possibly
3219 treated as an array symbol (for example C<@foo>),
3220 even though the same text in C<qq//> gives interpolation of C<\c@>.
3222 Code blocks such as C<(?{BLOCK})> are handled by temporarily passing control
3223 back to the perl parser, in a similar way that an interpolated array
3224 subscript expression such as C<"foo$array[1+f("[xyz")]bar"> would be.
3226 Moreover, inside C<(?{BLOCK})>, S<C<(?# comment )>>, and
3227 a C<#>-comment in a C</x>-regular expression, no processing is
3228 performed whatsoever. This is the first step at which the presence
3229 of the C</x> modifier is relevant.
3231 Interpolation in patterns has several quirks: C<$|>, C<$(>, C<$)>, C<@+>
3232 and C<@-> are not interpolated, and constructs C<$var[SOMETHING]> are
3233 voted (by several different estimators) to be either an array element
3234 or C<$var> followed by an RE alternative. This is where the notation
3235 C<${arr[$bar]}> comes handy: C</${arr[0-9]}/> is interpreted as
3236 array element C<-9>, not as a regular expression from the variable
3237 C<$arr> followed by a digit, which would be the interpretation of
3238 C</$arr[0-9]/>. Since voting among different estimators may occur,
3239 the result is not predictable.
3241 The lack of processing of C<\\> creates specific restrictions on
3242 the post-processed text. If the delimiter is C</>, one cannot get
3243 the combination C<\/> into the result of this step. C</> will
3244 finish the regular expression, C<\/> will be stripped to C</> on
3245 the previous step, and C<\\/> will be left as is. Because C</> is
3246 equivalent to C<\/> inside a regular expression, this does not
3247 matter unless the delimiter happens to be character special to the
3248 RE engine, such as in C<s*foo*bar*>, C<m[foo]>, or C<m?foo?>; or an
3249 alphanumeric char, as in:
3253 In the RE above, which is intentionally obfuscated for illustration, the
3254 delimiter is C<m>, the modifier is C<mx>, and after delimiter-removal the
3255 RE is the same as for S<C<m/ ^ a \s* b /mx>>. There's more than one
3256 reason you're encouraged to restrict your delimiters to non-alphanumeric,
3257 non-whitespace choices.
3261 This step is the last one for all constructs except regular expressions,
3262 which are processed further.
3264 =item Parsing regular expressions
3267 Previous steps were performed during the compilation of Perl code,
3268 but this one happens at run time, although it may be optimized to
3269 be calculated at compile time if appropriate. After preprocessing
3270 described above, and possibly after evaluation if concatenation,
3271 joining, casing translation, or metaquoting are involved, the
3272 resulting I<string> is passed to the RE engine for compilation.
3274 Whatever happens in the RE engine might be better discussed in L<perlre>,
3275 but for the sake of continuity, we shall do so here.
3277 This is another step where the presence of the C</x> modifier is
3278 relevant. The RE engine scans the string from left to right and
3279 converts it into a finite automaton.
3281 Backslashed characters are either replaced with corresponding
3282 literal strings (as with C<\{>), or else they generate special nodes
3283 in the finite automaton (as with C<\b>). Characters special to the
3284 RE engine (such as C<|>) generate corresponding nodes or groups of
3285 nodes. C<(?#...)> comments are ignored. All the rest is either
3286 converted to literal strings to match, or else is ignored (as is
3287 whitespace and C<#>-style comments if C</x> is present).
3289 Parsing of the bracketed character class construct, C<[...]>, is
3290 rather different than the rule used for the rest of the pattern.
3291 The terminator of this construct is found using the same rules as
3292 for finding the terminator of a C<{}>-delimited construct, the only
3293 exception being that C<]> immediately following C<[> is treated as
3294 though preceded by a backslash.
3296 The terminator of runtime C<(?{...})> is found by temporarily switching
3297 control to the perl parser, which should stop at the point where the
3298 logically balancing terminating C<}> is found.
3300 It is possible to inspect both the string given to RE engine and the
3301 resulting finite automaton. See the arguments C<debug>/C<debugcolor>
3302 in the S<C<use L<re>>> pragma, as well as Perl's B<-Dr> command-line
3303 switch documented in L<perlrun/"Command Switches">.
3305 =item Optimization of regular expressions
3306 X<regexp, optimization>
3308 This step is listed for completeness only. Since it does not change
3309 semantics, details of this step are not documented and are subject
3310 to change without notice. This step is performed over the finite
3311 automaton that was generated during the previous pass.
3313 It is at this stage that C<split()> silently optimizes C</^/> to
3318 =head2 I/O Operators
3319 X<operator, i/o> X<operator, io> X<io> X<while> X<filehandle>
3320 X<< <> >> X<< <<>> >> X<@ARGV>
3322 There are several I/O operators you should know about.
3324 A string enclosed by backticks (grave accents) first undergoes
3325 double-quote interpolation. It is then interpreted as an external
3326 command, and the output of that command is the value of the
3327 backtick string, like in a shell. In scalar context, a single string
3328 consisting of all output is returned. In list context, a list of
3329 values is returned, one per line of output. (You can set C<$/> to use
3330 a different line terminator.) The command is executed each time the
3331 pseudo-literal is evaluated. The status value of the command is
3332 returned in C<$?> (see L<perlvar> for the interpretation of C<$?>).
3333 Unlike in B<csh>, no translation is done on the return data--newlines
3334 remain newlines. Unlike in any of the shells, single quotes do not
3335 hide variable names in the command from interpretation. To pass a
3336 literal dollar-sign through to the shell you need to hide it with a
3337 backslash. The generalized form of backticks is C<qx//>, or you can
3338 call the L<perlfunc/readpipe> function. (Because
3339 backticks always undergo shell expansion as well, see L<perlsec> for
3341 X<qx> X<`> X<``> X<backtick> X<glob>
3343 In scalar context, evaluating a filehandle in angle brackets yields
3344 the next line from that file (the newline, if any, included), or
3345 C<undef> at end-of-file or on error. When C<$/> is set to C<undef>
3346 (sometimes known as file-slurp mode) and the file is empty, it
3347 returns C<''> the first time, followed by C<undef> subsequently.
3349 Ordinarily you must assign the returned value to a variable, but
3350 there is one situation where an automatic assignment happens. If
3351 and only if the input symbol is the only thing inside the conditional
3352 of a C<while> statement (even if disguised as a C<for(;;)> loop),
3353 the value is automatically assigned to the global variable C<$_>,
3354 destroying whatever was there previously. (This may seem like an
3355 odd thing to you, but you'll use the construct in almost every Perl
3356 script you write.) The C<$_> variable is not implicitly localized.
3357 You'll have to put a S<C<local $_;>> before the loop if you want that
3358 to happen. Furthermore, if the input symbol or an explicit assignment
3359 of the input symbol to a scalar is used as a C<while>/C<for> condition,
3360 then the condition actually tests for definedness of the expression's
3361 value, not for its regular truth value.
3363 Thus the following lines are equivalent:
3365 while (defined($_ = <STDIN>)) { print; }
3366 while ($_ = <STDIN>) { print; }
3367 while (<STDIN>) { print; }
3368 for (;<STDIN>;) { print; }
3369 print while defined($_ = <STDIN>);
3370 print while ($_ = <STDIN>);
3371 print while <STDIN>;
3373 This also behaves similarly, but assigns to a lexical variable
3374 instead of to C<$_>:
3376 while (my $line = <STDIN>) { print $line }
3378 In these loop constructs, the assigned value (whether assignment
3379 is automatic or explicit) is then tested to see whether it is
3380 defined. The defined test avoids problems where the line has a string
3381 value that would be treated as false by Perl; for example a "" or
3382 a C<"0"> with no trailing newline. If you really mean for such values
3383 to terminate the loop, they should be tested for explicitly:
3385 while (($_ = <STDIN>) ne '0') { ... }
3386 while (<STDIN>) { last unless $_; ... }
3388 In other boolean contexts, C<< <I<FILEHANDLE>> >> without an
3389 explicit C<defined> test or comparison elicits a warning if the
3390 S<C<use warnings>> pragma or the B<-w>
3391 command-line switch (the C<$^W> variable) is in effect.
3393 The filehandles STDIN, STDOUT, and STDERR are predefined. (The
3394 filehandles C<stdin>, C<stdout>, and C<stderr> will also work except
3395 in packages, where they would be interpreted as local identifiers
3396 rather than global.) Additional filehandles may be created with
3397 the C<open()> function, amongst others. See L<perlopentut> and
3398 L<perlfunc/open> for details on this.
3399 X<stdin> X<stdout> X<sterr>
3401 If a C<< <I<FILEHANDLE>> >> is used in a context that is looking for
3402 a list, a list comprising all input lines is returned, one line per
3403 list element. It's easy to grow to a rather large data space this
3404 way, so use with care.
3406 C<< <I<FILEHANDLE>> >> may also be spelled C<readline(*I<FILEHANDLE>)>.
3407 See L<perlfunc/readline>.
3409 The null filehandle C<< <> >> (sometimes called the diamond operator) is
3410 special: it can be used to emulate the
3411 behavior of B<sed> and B<awk>, and any other Unix filter program
3412 that takes a list of filenames, doing the same to each line
3413 of input from all of them. Input from C<< <> >> comes either from
3414 standard input, or from each file listed on the command line. Here's
3415 how it works: the first time C<< <> >> is evaluated, the C<@ARGV> array is
3416 checked, and if it is empty, C<$ARGV[0]> is set to C<"-">, which when opened
3417 gives you standard input. The C<@ARGV> array is then processed as a list
3418 of filenames. The loop
3421 ... # code for each line
3424 is equivalent to the following Perl-like pseudo code:
3426 unshift(@ARGV, '-') unless @ARGV;
3427 while ($ARGV = shift) {
3430 ... # code for each line
3434 except that it isn't so cumbersome to say, and will actually work.
3435 It really does shift the C<@ARGV> array and put the current filename
3436 into the C<$ARGV> variable. It also uses filehandle I<ARGV>
3437 internally. C<< <> >> is just a synonym for C<< <ARGV> >>, which
3438 is magical. (The pseudo code above doesn't work because it treats
3439 C<< <ARGV> >> as non-magical.)
3441 Since the null filehandle uses the two argument form of L<perlfunc/open>
3442 it interprets special characters, so if you have a script like this:
3448 and call it with S<C<perl dangerous.pl 'rm -rfv *|'>>, it actually opens a
3449 pipe, executes the C<rm> command and reads C<rm>'s output from that pipe.
3450 If you want all items in C<@ARGV> to be interpreted as file names, you
3451 can use the module C<ARGV::readonly> from CPAN, or use the double
3458 Using double angle brackets inside of a while causes the open to use the
3459 three argument form (with the second argument being C<< < >>), so all
3460 arguments in C<ARGV> are treated as literal filenames (including C<"-">).
3461 (Note that for convenience, if you use C<< <<>> >> and if C<@ARGV> is
3462 empty, it will still read from the standard input.)
3464 You can modify C<@ARGV> before the first C<< <> >> as long as the array ends up
3465 containing the list of filenames you really want. Line numbers (C<$.>)
3466 continue as though the input were one big happy file. See the example
3467 in L<perlfunc/eof> for how to reset line numbers on each file.
3469 If you want to set C<@ARGV> to your own list of files, go right ahead.
3470 This sets C<@ARGV> to all plain text files if no C<@ARGV> was given:
3472 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
3474 You can even set them to pipe commands. For example, this automatically
3475 filters compressed arguments through B<gzip>:
3477 @ARGV = map { /\.(gz|Z)$/ ? "gzip -dc < $_ |" : $_ } @ARGV;
3479 If you want to pass switches into your script, you can use one of the
3480 C<Getopts> modules or put a loop on the front like this:
3482 while ($_ = $ARGV[0], /^-/) {
3485 if (/^-D(.*)/) { $debug = $1 }
3486 if (/^-v/) { $verbose++ }
3487 # ... # other switches
3491 # ... # code for each line
3494 The C<< <> >> symbol will return C<undef> for end-of-file only once.
3495 If you call it again after this, it will assume you are processing another
3496 C<@ARGV> list, and if you haven't set C<@ARGV>, will read input from STDIN.
3498 If what the angle brackets contain is a simple scalar variable (for example,
3499 C<$foo>), then that variable contains the name of the
3500 filehandle to input from, or its typeglob, or a reference to the
3506 If what's within the angle brackets is neither a filehandle nor a simple
3507 scalar variable containing a filehandle name, typeglob, or typeglob
3508 reference, it is interpreted as a filename pattern to be globbed, and
3509 either a list of filenames or the next filename in the list is returned,
3510 depending on context. This distinction is determined on syntactic
3511 grounds alone. That means C<< <$x> >> is always a C<readline()> from
3512 an indirect handle, but C<< <$hash{key}> >> is always a C<glob()>.
3513 That's because C<$x> is a simple scalar variable, but C<$hash{key}> is
3514 not--it's a hash element. Even C<< <$x > >> (note the extra space)
3515 is treated as C<glob("$x ")>, not C<readline($x)>.
3517 One level of double-quote interpretation is done first, but you can't
3518 say C<< <$foo> >> because that's an indirect filehandle as explained
3519 in the previous paragraph. (In older versions of Perl, programmers
3520 would insert curly brackets to force interpretation as a filename glob:
3521 C<< <${foo}> >>. These days, it's considered cleaner to call the
3522 internal function directly as C<glob($foo)>, which is probably the right
3523 way to have done it in the first place.) For example:
3529 is roughly equivalent to:
3531 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
3537 except that the globbing is actually done internally using the standard
3538 C<L<File::Glob>> extension. Of course, the shortest way to do the above is:
3542 A (file)glob evaluates its (embedded) argument only when it is
3543 starting a new list. All values must be read before it will start
3544 over. In list context, this isn't important because you automatically
3545 get them all anyway. However, in scalar context the operator returns
3546 the next value each time it's called, or C<undef> when the list has
3547 run out. As with filehandle reads, an automatic C<defined> is
3548 generated when the glob occurs in the test part of a C<while>,
3549 because legal glob returns (for example,
3550 a file called F<0>) would otherwise
3551 terminate the loop. Again, C<undef> is returned only once. So if
3552 you're expecting a single value from a glob, it is much better to
3555 ($file) = <blurch*>;
3561 because the latter will alternate between returning a filename and
3564 If you're trying to do variable interpolation, it's definitely better
3565 to use the C<glob()> function, because the older notation can cause people
3566 to become confused with the indirect filehandle notation.
3568 @files = glob("$dir/*.[ch]");
3569 @files = glob($files[$i]);
3571 If an angle-bracket-based globbing expression is used as the condition of
3572 a C<while> or C<for> loop, then it will be implicitly assigned to C<$_>.
3573 If either a globbing expression or an explicit assignment of a globbing
3574 expression to a scalar is used as a C<while>/C<for> condition, then
3575 the condition actually tests for definedness of the expression's value,
3576 not for its regular truth value.
3578 =head2 Constant Folding
3579 X<constant folding> X<folding>
3581 Like C, Perl does a certain amount of expression evaluation at
3582 compile time whenever it determines that all arguments to an
3583 operator are static and have no side effects. In particular, string
3584 concatenation happens at compile time between literals that don't do
3585 variable substitution. Backslash interpolation also happens at
3586 compile time. You can say
3588 'Now is the time for all'
3590 . 'good men to come to.'
3592 and this all reduces to one string internally. Likewise, if
3595 foreach $file (@filenames) {
3596 if (-s $file > 5 + 100 * 2**16) { }
3599 the compiler precomputes the number which that expression
3600 represents so that the interpreter won't have to.
3605 Perl doesn't officially have a no-op operator, but the bare constants
3606 C<0> and C<1> are special-cased not to produce a warning in void
3607 context, so you can for example safely do
3611 =head2 Bitwise String Operators
3612 X<operator, bitwise, string> X<&.> X<|.> X<^.> X<~.>
3614 Bitstrings of any size may be manipulated by the bitwise operators
3617 If the operands to a binary bitwise op are strings of different
3618 sizes, B<|> and B<^> ops act as though the shorter operand had
3619 additional zero bits on the right, while the B<&> op acts as though
3620 the longer operand were truncated to the length of the shorter.
3621 The granularity for such extension or truncation is one or more
3624 # ASCII-based examples
3625 print "j p \n" ^ " a h"; # prints "JAPH\n"
3626 print "JA" | " ph\n"; # prints "japh\n"
3627 print "japh\nJunk" & '_____'; # prints "JAPH\n";
3628 print 'p N$' ^ " E<H\n"; # prints "Perl\n";
3630 If you are intending to manipulate bitstrings, be certain that
3631 you're supplying bitstrings: If an operand is a number, that will imply
3632 a B<numeric> bitwise operation. You may explicitly show which type of
3633 operation you intend by using C<""> or C<0+>, as in the examples below.
3635 $foo = 150 | 105; # yields 255 (0x96 | 0x69 is 0xFF)
3636 $foo = '150' | 105; # yields 255
3637 $foo = 150 | '105'; # yields 255
3638 $foo = '150' | '105'; # yields string '155' (under ASCII)
3640 $baz = 0+$foo & 0+$bar; # both ops explicitly numeric
3641 $biz = "$foo" ^ "$bar"; # both ops explicitly stringy
3643 This somewhat unpredictable behavior can be avoided with the "bitwise"
3644 feature, new in Perl 5.22. You can enable it via S<C<use feature
3645 'bitwise'>> or C<use v5.28>. Before Perl 5.28, it used to emit a warning
3646 in the C<"experimental::bitwise"> category. Under this feature, the four
3647 standard bitwise operators (C<~ | & ^>) are always numeric. Adding a dot
3648 after each operator (C<~. |. &. ^.>) forces it to treat its operands as
3651 use feature "bitwise";
3652 $foo = 150 | 105; # yields 255 (0x96 | 0x69 is 0xFF)
3653 $foo = '150' | 105; # yields 255
3654 $foo = 150 | '105'; # yields 255
3655 $foo = '150' | '105'; # yields 255
3656 $foo = 150 |. 105; # yields string '155'
3657 $foo = '150' |. 105; # yields string '155'
3658 $foo = 150 |.'105'; # yields string '155'
3659 $foo = '150' |.'105'; # yields string '155'
3661 $baz = $foo & $bar; # both operands numeric
3662 $biz = $foo ^. $bar; # both operands stringy
3664 The assignment variants of these operators (C<&= |= ^= &.= |.= ^.=>)
3665 behave likewise under the feature.
3667 It is a fatal error if an operand contains a character whose ordinal
3668 value is above 0xFF, and hence not expressible except in UTF-8. The
3669 operation is performed on a non-UTF-8 copy for other operands encoded in
3670 UTF-8. See L<perlunicode/Byte and Character Semantics>.
3672 See L<perlfunc/vec> for information on how to manipulate individual bits
3675 =head2 Integer Arithmetic
3678 By default, Perl assumes that it must do most of its arithmetic in
3679 floating point. But by saying
3683 you may tell the compiler to use integer operations
3684 (see L<integer> for a detailed explanation) from here to the end of
3685 the enclosing BLOCK. An inner BLOCK may countermand this by saying
3689 which lasts until the end of that BLOCK. Note that this doesn't
3690 mean everything is an integer, merely that Perl will use integer
3691 operations for arithmetic, comparison, and bitwise operators. For
3692 example, even under S<C<use integer>>, if you take the C<sqrt(2)>, you'll
3693 still get C<1.4142135623731> or so.
3695 Used on numbers, the bitwise operators (C<&> C<|> C<^> C<~> C<< << >>
3696 C<< >> >>) always produce integral results. (But see also
3697 L</Bitwise String Operators>.) However, S<C<use integer>> still has meaning for
3698 them. By default, their results are interpreted as unsigned integers, but
3699 if S<C<use integer>> is in effect, their results are interpreted
3700 as signed integers. For example, C<~0> usually evaluates to a large
3701 integral value. However, S<C<use integer; ~0>> is C<-1> on two's-complement
3704 =head2 Floating-point Arithmetic
3706 X<floating-point> X<floating point> X<float> X<real>
3708 While S<C<use integer>> provides integer-only arithmetic, there is no
3709 analogous mechanism to provide automatic rounding or truncation to a
3710 certain number of decimal places. For rounding to a certain number
3711 of digits, C<sprintf()> or C<printf()> is usually the easiest route.
3714 Floating-point numbers are only approximations to what a mathematician
3715 would call real numbers. There are infinitely more reals than floats,
3716 so some corners must be cut. For example:
3718 printf "%.20g\n", 123456789123456789;
3719 # produces 123456789123456784
3721 Testing for exact floating-point equality or inequality is not a
3722 good idea. Here's a (relatively expensive) work-around to compare
3723 whether two floating-point numbers are equal to a particular number of
3724 decimal places. See Knuth, volume II, for a more robust treatment of
3728 my ($X, $Y, $POINTS) = @_;
3730 $tX = sprintf("%.${POINTS}g", $X);
3731 $tY = sprintf("%.${POINTS}g", $Y);
3735 The POSIX module (part of the standard perl distribution) implements
3736 C<ceil()>, C<floor()>, and other mathematical and trigonometric functions.
3737 The C<L<Math::Complex>> module (part of the standard perl distribution)
3738 defines mathematical functions that work on both the reals and the
3739 imaginary numbers. C<Math::Complex> is not as efficient as POSIX, but
3740 POSIX can't work with complex numbers.
3742 Rounding in financial applications can have serious implications, and
3743 the rounding method used should be specified precisely. In these
3744 cases, it probably pays not to trust whichever system rounding is
3745 being used by Perl, but to instead implement the rounding function you
3748 =head2 Bigger Numbers
3749 X<number, arbitrary precision>
3751 The standard C<L<Math::BigInt>>, C<L<Math::BigRat>>, and
3752 C<L<Math::BigFloat>> modules,
3753 along with the C<bignum>, C<bigint>, and C<bigrat> pragmas, provide
3754 variable-precision arithmetic and overloaded operators, although
3755 they're currently pretty slow. At the cost of some space and
3756 considerable speed, they avoid the normal pitfalls associated with
3757 limited-precision representations.
3760 use bigint; # easy interface to Math::BigInt
3761 $x = 123456789123456789;
3763 +15241578780673678515622620750190521
3771 say "x/y is ", $x/$y;
3772 say "x*y is ", $x*$y;
3776 Several modules let you calculate with unlimited or fixed precision
3777 (bound only by memory and CPU time). There
3778 are also some non-standard modules that
3779 provide faster implementations via external C libraries.
3781 Here is a short, but incomplete summary:
3783 Math::String treat string sequences like numbers
3784 Math::FixedPrecision calculate with a fixed precision
3785 Math::Currency for currency calculations
3786 Bit::Vector manipulate bit vectors fast (uses C)
3787 Math::BigIntFast Bit::Vector wrapper for big numbers
3788 Math::Pari provides access to the Pari C library
3789 Math::Cephes uses the external Cephes C library (no
3791 Math::Cephes::Fraction fractions via the Cephes library
3792 Math::GMP another one using an external C library
3793 Math::GMPz an alternative interface to libgmp's big ints
3794 Math::GMPq an interface to libgmp's fraction numbers
3795 Math::GMPf an interface to libgmp's floating point numbers