3 perlop - Perl operators and precedence
7 Perl operators have the following associativity and precedence,
8 listed from highest precedence to lowest. Operators borrowed from
9 C keep the same precedence relationship with each other, even where
10 C's precedence is slightly screwy. (This makes learning Perl easier
11 for C folks.) With very few exceptions, these all operate on scalar
12 values only, not array values.
14 left terms and list operators (leftward)
18 right ! ~ \ and unary + and -
23 nonassoc named unary operators
24 nonassoc < > <= >= lt gt le ge
25 nonassoc == != <=> eq ne cmp
34 nonassoc list operators (rightward)
39 In the following sections, these operators are covered in precedence order.
41 Many operators can be overloaded for objects. See L<overload>.
45 =head2 Terms and List Operators (Leftward)
47 A TERM has the highest precedence in Perl. They include variables,
48 quote and quote-like operators, any expression in parentheses,
49 and any function whose arguments are parenthesized. Actually, there
50 aren't really functions in this sense, just list operators and unary
51 operators behaving as functions because you put parentheses around
52 the arguments. These are all documented in L<perlfunc>.
54 If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
55 is followed by a left parenthesis as the next token, the operator and
56 arguments within parentheses are taken to be of highest precedence,
57 just like a normal function call.
59 In the absence of parentheses, the precedence of list operators such as
60 C<print>, C<sort>, or C<chmod> is either very high or very low depending on
61 whether you are looking at the left side or the right side of the operator.
64 @ary = (1, 3, sort 4, 2);
65 print @ary; # prints 1324
67 the commas on the right of the sort are evaluated before the sort,
68 but the commas on the left are evaluated after. In other words,
69 list operators tend to gobble up all arguments that follow, and
70 then act like a simple TERM with regard to the preceding expression.
71 Be careful with parentheses:
73 # These evaluate exit before doing the print:
74 print($foo, exit); # Obviously not what you want.
75 print $foo, exit; # Nor is this.
77 # These do the print before evaluating exit:
78 (print $foo), exit; # This is what you want.
79 print($foo), exit; # Or this.
80 print ($foo), exit; # Or even this.
84 print ($foo & 255) + 1, "\n";
86 probably doesn't do what you expect at first glance. See
87 L<Named Unary Operators> for more discussion of this.
89 Also parsed as terms are the C<do {}> and C<eval {}> constructs, as
90 well as subroutine and method calls, and the anonymous
91 constructors C<[]> and C<{}>.
93 See also L<Quote and Quote-like Operators> toward the end of this section,
94 as well as L<"I/O Operators">.
96 =head2 The Arrow Operator
98 "C<-E<gt>>" is an infix dereference operator, just as it is in C
99 and C++. If the right side is either a C<[...]>, C<{...}>, or a
100 C<(...)> subscript, then the left side must be either a hard or
101 symbolic reference to an array, a hash, or a subroutine respectively.
102 (Or technically speaking, a location capable of holding a hard
103 reference, if it's an array or hash reference being used for
104 assignment.) See L<perlreftut> and L<perlref>.
106 Otherwise, the right side is a method name or a simple scalar
107 variable containing either the method name or a subroutine reference,
108 and the left side must be either an object (a blessed reference)
109 or a class name (that is, a package name). See L<perlobj>.
111 =head2 Auto-increment and Auto-decrement
113 "++" and "--" work as in C. That is, if placed before a variable, they
114 increment or decrement the variable before returning the value, and if
115 placed after, increment or decrement the variable after returning the value.
117 The auto-increment operator has a little extra builtin magic to it. If
118 you increment a variable that is numeric, or that has ever been used in
119 a numeric context, you get a normal increment. If, however, the
120 variable has been used in only string contexts since it was set, and
121 has a value that is not the empty string and matches the pattern
122 C</^[a-zA-Z]*[0-9]*$/>, the increment is done as a string, preserving each
123 character within its range, with carry:
125 print ++($foo = '99'); # prints '100'
126 print ++($foo = 'a0'); # prints 'a1'
127 print ++($foo = 'Az'); # prints 'Ba'
128 print ++($foo = 'zz'); # prints 'aaa'
130 The auto-decrement operator is not magical.
132 =head2 Exponentiation
134 Binary "**" is the exponentiation operator. It binds even more
135 tightly than unary minus, so -2**4 is -(2**4), not (-2)**4. (This is
136 implemented using C's pow(3) function, which actually works on doubles
139 =head2 Symbolic Unary Operators
141 Unary "!" performs logical negation, i.e., "not". See also C<not> for a lower
142 precedence version of this.
144 Unary "-" performs arithmetic negation if the operand is numeric. If
145 the operand is an identifier, a string consisting of a minus sign
146 concatenated with the identifier is returned. Otherwise, if the string
147 starts with a plus or minus, a string starting with the opposite sign
148 is returned. One effect of these rules is that C<-bareword> is equivalent
151 Unary "~" performs bitwise negation, i.e., 1's complement. For example,
152 C<0666 &~ 027> is 0640. (See also L<Integer Arithmetic> and L<Bitwise
155 Unary "+" has no effect whatsoever, even on strings. It is useful
156 syntactically for separating a function name from a parenthesized expression
157 that would otherwise be interpreted as the complete list of function
158 arguments. (See examples above under L<Terms and List Operators (Leftward)>.)
160 Unary "\" creates a reference to whatever follows it. See L<perlreftut>
161 and L<perlref>. Do not confuse this behavior with the behavior of
162 backslash within a string, although both forms do convey the notion
163 of protecting the next thing from interpolation.
165 =head2 Binding Operators
167 Binary "=~" binds a scalar expression to a pattern match. Certain operations
168 search or modify the string $_ by default. This operator makes that kind
169 of operation work on some other string. The right argument is a search
170 pattern, substitution, or transliteration. The left argument is what is
171 supposed to be searched, substituted, or transliterated instead of the default
172 $_. The return value indicates the success of the operation. (If the
173 right argument is an expression rather than a search pattern,
174 substitution, or transliteration, it is interpreted as a search pattern at run
175 time. This can be is less efficient than an explicit search, because the
176 pattern must be compiled every time the expression is evaluated.
178 Binary "!~" is just like "=~" except the return value is negated in
181 =head2 Multiplicative Operators
183 Binary "*" multiplies two numbers.
185 Binary "/" divides two numbers.
187 Binary "%" computes the modulus of two numbers. Given integer
188 operands C<$a> and C<$b>: If C<$b> is positive, then C<$a % $b> is
189 C<$a> minus the largest multiple of C<$b> that is not greater than
190 C<$a>. If C<$b> is negative, then C<$a % $b> is C<$a> minus the
191 smallest multiple of C<$b> that is not less than C<$a> (i.e. the
192 result will be less than or equal to zero).
193 Note than when C<use integer> is in scope, "%" give you direct access
194 to the modulus operator as implemented by your C compiler. This
195 operator is not as well defined for negative operands, but it will
198 Binary "x" is the repetition operator. In scalar context, it
199 returns a string consisting of the left operand repeated the number of
200 times specified by the right operand. In list context, if the left
201 operand is a list in parentheses, it repeats the list.
203 print '-' x 80; # print row of dashes
205 print "\t" x ($tab/8), ' ' x ($tab%8); # tab over
207 @ones = (1) x 80; # a list of 80 1's
208 @ones = (5) x @ones; # set all elements to 5
211 =head2 Additive Operators
213 Binary "+" returns the sum of two numbers.
215 Binary "-" returns the difference of two numbers.
217 Binary "." concatenates two strings.
219 =head2 Shift Operators
221 Binary "<<" returns the value of its left argument shifted left by the
222 number of bits specified by the right argument. Arguments should be
223 integers. (See also L<Integer Arithmetic>.)
225 Binary ">>" returns the value of its left argument shifted right by
226 the number of bits specified by the right argument. Arguments should
227 be integers. (See also L<Integer Arithmetic>.)
229 =head2 Named Unary Operators
231 The various named unary operators are treated as functions with one
232 argument, with optional parentheses. These include the filetest
233 operators, like C<-f>, C<-M>, etc. See L<perlfunc>.
235 If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
236 is followed by a left parenthesis as the next token, the operator and
237 arguments within parentheses are taken to be of highest precedence,
238 just like a normal function call. Examples:
240 chdir $foo || die; # (chdir $foo) || die
241 chdir($foo) || die; # (chdir $foo) || die
242 chdir ($foo) || die; # (chdir $foo) || die
243 chdir +($foo) || die; # (chdir $foo) || die
245 but, because * is higher precedence than ||:
247 chdir $foo * 20; # chdir ($foo * 20)
248 chdir($foo) * 20; # (chdir $foo) * 20
249 chdir ($foo) * 20; # (chdir $foo) * 20
250 chdir +($foo) * 20; # chdir ($foo * 20)
252 rand 10 * 20; # rand (10 * 20)
253 rand(10) * 20; # (rand 10) * 20
254 rand (10) * 20; # (rand 10) * 20
255 rand +(10) * 20; # rand (10 * 20)
257 See also L<"Terms and List Operators (Leftward)">.
259 =head2 Relational Operators
261 Binary "E<lt>" returns true if the left argument is numerically less than
264 Binary "E<gt>" returns true if the left argument is numerically greater
265 than the right argument.
267 Binary "E<lt>=" returns true if the left argument is numerically less than
268 or equal to the right argument.
270 Binary "E<gt>=" returns true if the left argument is numerically greater
271 than or equal to the right argument.
273 Binary "lt" returns true if the left argument is stringwise less than
276 Binary "gt" returns true if the left argument is stringwise greater
277 than the right argument.
279 Binary "le" returns true if the left argument is stringwise less than
280 or equal to the right argument.
282 Binary "ge" returns true if the left argument is stringwise greater
283 than or equal to the right argument.
285 =head2 Equality Operators
287 Binary "==" returns true if the left argument is numerically equal to
290 Binary "!=" returns true if the left argument is numerically not equal
291 to the right argument.
293 Binary "E<lt>=E<gt>" returns -1, 0, or 1 depending on whether the left
294 argument is numerically less than, equal to, or greater than the right
297 Binary "eq" returns true if the left argument is stringwise equal to
300 Binary "ne" returns true if the left argument is stringwise not equal
301 to the right argument.
303 Binary "cmp" returns -1, 0, or 1 depending on whether the left argument is stringwise
304 less than, equal to, or greater than the right argument.
306 "lt", "le", "ge", "gt" and "cmp" use the collation (sort) order specified
307 by the current locale if C<use locale> is in effect. See L<perllocale>.
311 Binary "&" returns its operators ANDed together bit by bit.
312 (See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
314 =head2 Bitwise Or and Exclusive Or
316 Binary "|" returns its operators ORed together bit by bit.
317 (See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
319 Binary "^" returns its operators XORed together bit by bit.
320 (See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
322 =head2 C-style Logical And
324 Binary "&&" performs a short-circuit logical AND operation. That is,
325 if the left operand is false, the right operand is not even evaluated.
326 Scalar or list context propagates down to the right operand if it
329 =head2 C-style Logical Or
331 Binary "||" performs a short-circuit logical OR operation. That is,
332 if the left operand is true, the right operand is not even evaluated.
333 Scalar or list context propagates down to the right operand if it
336 The C<||> and C<&&> operators differ from C's in that, rather than returning
337 0 or 1, they return the last value evaluated. Thus, a reasonably portable
338 way to find out the home directory (assuming it's not "0") might be:
340 $home = $ENV{'HOME'} || $ENV{'LOGDIR'} ||
341 (getpwuid($<))[7] || die "You're homeless!\n";
343 In particular, this means that you shouldn't use this
344 for selecting between two aggregates for assignment:
346 @a = @b || @c; # this is wrong
347 @a = scalar(@b) || @c; # really meant this
348 @a = @b ? @b : @c; # this works fine, though
350 As more readable alternatives to C<&&> and C<||> when used for
351 control flow, Perl provides C<and> and C<or> operators (see below).
352 The short-circuit behavior is identical. The precedence of "and" and
353 "or" is much lower, however, so that you can safely use them after a
354 list operator without the need for parentheses:
356 unlink "alpha", "beta", "gamma"
357 or gripe(), next LINE;
359 With the C-style operators that would have been written like this:
361 unlink("alpha", "beta", "gamma")
362 || (gripe(), next LINE);
364 Use "or" for assignment is unlikely to do what you want; see below.
366 =head2 Range Operators
368 Binary ".." is the range operator, which is really two different
369 operators depending on the context. In list context, it returns an
370 array of values counting (up by ones) from the left value to the right
371 value. If the left value is greater than the right value then it
372 returns the empty array. The range operator is useful for writing
373 C<foreach (1..10)> loops and for doing slice operations on arrays. In
374 the current implementation, no temporary array is created when the
375 range operator is used as the expression in C<foreach> loops, but older
376 versions of Perl might burn a lot of memory when you write something
379 for (1 .. 1_000_000) {
383 In scalar context, ".." returns a boolean value. The operator is
384 bistable, like a flip-flop, and emulates the line-range (comma) operator
385 of B<sed>, B<awk>, and various editors. Each ".." operator maintains its
386 own boolean state. It is false as long as its left operand is false.
387 Once the left operand is true, the range operator stays true until the
388 right operand is true, I<AFTER> which the range operator becomes false
389 again. It doesn't become false till the next time the range operator is
390 evaluated. It can test the right operand and become false on the same
391 evaluation it became true (as in B<awk>), but it still returns true once.
392 If you don't want it to test the right operand till the next
393 evaluation, as in B<sed>, just use three dots ("...") instead of
394 two. In all other regards, "..." behaves just like ".." does.
396 The right operand is not evaluated while the operator is in the
397 "false" state, and the left operand is not evaluated while the
398 operator is in the "true" state. The precedence is a little lower
399 than || and &&. The value returned is either the empty string for
400 false, or a sequence number (beginning with 1) for true. The
401 sequence number is reset for each range encountered. The final
402 sequence number in a range has the string "E0" appended to it, which
403 doesn't affect its numeric value, but gives you something to search
404 for if you want to exclude the endpoint. You can exclude the
405 beginning point by waiting for the sequence number to be greater
406 than 1. If either operand of scalar ".." is a constant expression,
407 that operand is implicitly compared to the C<$.> variable, the
408 current line number. Examples:
410 As a scalar operator:
412 if (101 .. 200) { print; } # print 2nd hundred lines
413 next line if (1 .. /^$/); # skip header lines
414 s/^/> / if (/^$/ .. eof()); # quote body
416 # parse mail messages
418 $in_header = 1 .. /^$/;
419 $in_body = /^$/ .. eof();
420 # do something based on those
422 close ARGV if eof; # reset $. each file
427 for (101 .. 200) { print; } # print $_ 100 times
428 @foo = @foo[0 .. $#foo]; # an expensive no-op
429 @foo = @foo[$#foo-4 .. $#foo]; # slice last 5 items
431 The range operator (in list context) makes use of the magical
432 auto-increment algorithm if the operands are strings. You
435 @alphabet = ('A' .. 'Z');
437 to get all normal letters of the alphabet, or
439 $hexdigit = (0 .. 9, 'a' .. 'f')[$num & 15];
441 to get a hexadecimal digit, or
443 @z2 = ('01' .. '31'); print $z2[$mday];
445 to get dates with leading zeros. If the final value specified is not
446 in the sequence that the magical increment would produce, the sequence
447 goes until the next value would be longer than the final value
450 =head2 Conditional Operator
452 Ternary "?:" is the conditional operator, just as in C. It works much
453 like an if-then-else. If the argument before the ? is true, the
454 argument before the : is returned, otherwise the argument after the :
455 is returned. For example:
457 printf "I have %d dog%s.\n", $n,
458 ($n == 1) ? '' : "s";
460 Scalar or list context propagates downward into the 2nd
461 or 3rd argument, whichever is selected.
463 $a = $ok ? $b : $c; # get a scalar
464 @a = $ok ? @b : @c; # get an array
465 $a = $ok ? @b : @c; # oops, that's just a count!
467 The operator may be assigned to if both the 2nd and 3rd arguments are
468 legal lvalues (meaning that you can assign to them):
470 ($a_or_b ? $a : $b) = $c;
472 Because this operator produces an assignable result, using assignments
473 without parentheses will get you in trouble. For example, this:
475 $a % 2 ? $a += 10 : $a += 2
479 (($a % 2) ? ($a += 10) : $a) += 2
483 ($a % 2) ? ($a += 10) : ($a += 2)
485 That should probably be written more simply as:
487 $a += ($a % 2) ? 10 : 2;
489 =head2 Assignment Operators
491 "=" is the ordinary assignment operator.
493 Assignment operators work as in C. That is,
501 although without duplicating any side effects that dereferencing the lvalue
502 might trigger, such as from tie(). Other assignment operators work similarly.
503 The following are recognized:
510 Although these are grouped by family, they all have the precedence
513 Unlike in C, the assignment operator produces a valid lvalue. Modifying
514 an assignment is equivalent to doing the assignment and then modifying
515 the variable that was assigned to. This is useful for modifying
516 a copy of something, like this:
518 ($tmp = $global) =~ tr [A-Z] [a-z];
529 =head2 Comma Operator
531 Binary "," is the comma operator. In scalar context it evaluates
532 its left argument, throws that value away, then evaluates its right
533 argument and returns that value. This is just like C's comma operator.
535 In list context, it's just the list argument separator, and inserts
536 both its arguments into the list.
538 The =E<gt> digraph is mostly just a synonym for the comma operator. It's useful for
539 documenting arguments that come in pairs. As of release 5.001, it also forces
540 any word to the left of it to be interpreted as a string.
542 =head2 List Operators (Rightward)
544 On the right side of a list operator, it has very low precedence,
545 such that it controls all comma-separated expressions found there.
546 The only operators with lower precedence are the logical operators
547 "and", "or", and "not", which may be used to evaluate calls to list
548 operators without the need for extra parentheses:
550 open HANDLE, "filename"
551 or die "Can't open: $!\n";
553 See also discussion of list operators in L<Terms and List Operators (Leftward)>.
557 Unary "not" returns the logical negation of the expression to its right.
558 It's the equivalent of "!" except for the very low precedence.
562 Binary "and" returns the logical conjunction of the two surrounding
563 expressions. It's equivalent to && except for the very low
564 precedence. This means that it short-circuits: i.e., the right
565 expression is evaluated only if the left expression is true.
567 =head2 Logical or and Exclusive Or
569 Binary "or" returns the logical disjunction of the two surrounding
570 expressions. It's equivalent to || except for the very low precedence.
571 This makes it useful for control flow
573 print FH $data or die "Can't write to FH: $!";
575 This means that it short-circuits: i.e., the right expression is evaluated
576 only if the left expression is false. Due to its precedence, you should
577 probably avoid using this for assignment, only for control flow.
579 $a = $b or $c; # bug: this is wrong
580 ($a = $b) or $c; # really means this
581 $a = $b || $c; # better written this way
583 However, when it's a list-context assignment and you're trying to use
584 "||" for control flow, you probably need "or" so that the assignment
585 takes higher precedence.
587 @info = stat($file) || die; # oops, scalar sense of stat!
588 @info = stat($file) or die; # better, now @info gets its due
590 Then again, you could always use parentheses.
592 Binary "xor" returns the exclusive-OR of the two surrounding expressions.
593 It cannot short circuit, of course.
595 =head2 C Operators Missing From Perl
597 Here is what C has that Perl doesn't:
603 Address-of operator. (But see the "\" operator for taking a reference.)
607 Dereference-address operator. (Perl's prefix dereferencing
608 operators are typed: $, @, %, and &.)
612 Type-casting operator.
616 =head2 Quote and Quote-like Operators
618 While we usually think of quotes as literal values, in Perl they
619 function as operators, providing various kinds of interpolating and
620 pattern matching capabilities. Perl provides customary quote characters
621 for these behaviors, but also provides a way for you to choose your
622 quote character for any of them. In the following table, a C<{}> represents
623 any pair of delimiters you choose.
625 Customary Generic Meaning Interpolates
628 `` qx{} Command yes (unless '' is delimiter)
630 // m{} Pattern match yes (unless '' is delimiter)
631 qr{} Pattern yes (unless '' is delimiter)
632 s{}{} Substitution yes (unless '' is delimiter)
633 tr{}{} Transliteration no (but see below)
635 Non-bracketing delimiters use the same character fore and aft, but the four
636 sorts of brackets (round, angle, square, curly) will all nest, which means
645 Note, however, that this does not always work for quoting Perl code:
647 $s = q{ if($a eq "}") ... }; # WRONG
649 is a syntax error. The C<Text::Balanced> module on CPAN is able to do this
652 There can be whitespace between the operator and the quoting
653 characters, except when C<#> is being used as the quoting character.
654 C<q#foo#> is parsed as the string C<foo>, while C<q #foo#> is the
655 operator C<q> followed by a comment. Its argument will be taken
656 from the next line. This allows you to write:
658 s {foo} # Replace foo
661 For constructs that do interpolate, variables beginning with "C<$>"
662 or "C<@>" are interpolated, as are the following escape sequences. Within
663 a transliteration, the first eleven of these sequences may be used.
670 \a alarm (bell) (BEL)
672 \033 octal char (ESC)
674 \x{263a} wide hex char (SMILEY)
675 \c[ control char (ESC)
677 \l lowercase next char
678 \u uppercase next char
681 \E end case modification
682 \Q quote non-word characters till \E
684 If C<use locale> is in effect, the case map used by C<\l>, C<\L>, C<\u>
685 and C<\U> is taken from the current locale. See L<perllocale>.
687 All systems use the virtual C<"\n"> to represent a line terminator,
688 called a "newline". There is no such thing as an unvarying, physical
689 newline character. It is only an illusion that the operating system,
690 device drivers, C libraries, and Perl all conspire to preserve. Not all
691 systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
692 on a Mac, these are reversed, and on systems without line terminator,
693 printing C<"\n"> may emit no actual data. In general, use C<"\n"> when
694 you mean a "newline" for your system, but use the literal ASCII when you
695 need an exact character. For example, most networking protocols expect
696 and prefer a CR+LF (C<"\012\015"> or C<"\cJ\cM">) for line terminators,
697 and although they often accept just C<"\012">, they seldom tolerate just
698 C<"\015">. If you get in the habit of using C<"\n"> for networking,
699 you may be burned some day.
701 You cannot include a literal C<$> or C<@> within a C<\Q> sequence.
702 An unescaped C<$> or C<@> interpolates the corresponding variable,
703 while escaping will cause the literal string C<\$> to be inserted.
704 You'll need to write something like C<m/\Quser\E\@\Qhost/>.
706 Patterns are subject to an additional level of interpretation as a
707 regular expression. This is done as a second pass, after variables are
708 interpolated, so that regular expressions may be incorporated into the
709 pattern from the variables. If this is not what you want, use C<\Q> to
710 interpolate a variable literally.
712 Apart from the behavior described above, Perl does not expand
713 multiple levels of interpolation. In particular, contrary to the
714 expectations of shell programmers, back-quotes do I<NOT> interpolate
715 within double quotes, nor do single quotes impede evaluation of
716 variables when used within double quotes.
718 =head2 Regexp Quote-Like Operators
720 Here are the quote-like operators that apply to pattern
721 matching and related activities.
727 This is just like the C</pattern/> search, except that it matches only
728 once between calls to the reset() operator. This is a useful
729 optimization when you want to see only the first occurrence of
730 something in each file of a set of files, for instance. Only C<??>
731 patterns local to the current package are reset.
735 # blank line between header and body
738 reset if eof; # clear ?? status for next file
741 This usage is vaguely depreciated, which means it just might possibly
742 be removed in some distant future version of Perl, perhaps somewhere
743 around the year 2168.
745 =item m/PATTERN/cgimosx
747 =item /PATTERN/cgimosx
749 Searches a string for a pattern match, and in scalar context returns
750 true if it succeeds, false if it fails. If no string is specified
751 via the C<=~> or C<!~> operator, the $_ string is searched. (The
752 string specified with C<=~> need not be an lvalue--it may be the
753 result of an expression evaluation, but remember the C<=~> binds
754 rather tightly.) See also L<perlre>. See L<perllocale> for
755 discussion of additional considerations that apply when C<use locale>
760 c Do not reset search position on a failed match when /g is in effect.
761 g Match globally, i.e., find all occurrences.
762 i Do case-insensitive pattern matching.
763 m Treat string as multiple lines.
764 o Compile pattern only once.
765 s Treat string as single line.
766 x Use extended regular expressions.
768 If "/" is the delimiter then the initial C<m> is optional. With the C<m>
769 you can use any pair of non-alphanumeric, non-whitespace characters
770 as delimiters. This is particularly useful for matching path names
771 that contain "/", to avoid LTS (leaning toothpick syndrome). If "?" is
772 the delimiter, then the match-only-once rule of C<?PATTERN?> applies.
773 If "'" is the delimiter, no interpolation is performed on the PATTERN.
775 PATTERN may contain variables, which will be interpolated (and the
776 pattern recompiled) every time the pattern search is evaluated, except
777 for when the delimiter is a single quote. (Note that C<$)> and C<$|>
778 might not be interpolated because they look like end-of-string tests.)
779 If you want such a pattern to be compiled only once, add a C</o> after
780 the trailing delimiter. This avoids expensive run-time recompilations,
781 and is useful when the value you are interpolating won't change over
782 the life of the script. However, mentioning C</o> constitutes a promise
783 that you won't change the variables in the pattern. If you change them,
784 Perl won't even notice. See also L<qr//>.
786 If the PATTERN evaluates to the empty string, the last
787 I<successfully> matched regular expression is used instead.
789 If the C</g> option is not used, C<m//> in list context returns a
790 list consisting of the subexpressions matched by the parentheses in the
791 pattern, i.e., (C<$1>, C<$2>, C<$3>...). (Note that here C<$1> etc. are
792 also set, and that this differs from Perl 4's behavior.) When there are
793 no parentheses in the pattern, the return value is the list C<(1)> for
794 success. With or without parentheses, an empty list is returned upon
799 open(TTY, '/dev/tty');
800 <TTY> =~ /^y/i && foo(); # do foo if desired
802 if (/Version: *([0-9.]*)/) { $version = $1; }
804 next if m#^/usr/spool/uucp#;
809 print if /$arg/o; # compile only once
812 if (($F1, $F2, $Etc) = ($foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
814 This last example splits $foo into the first two words and the
815 remainder of the line, and assigns those three fields to $F1, $F2, and
816 $Etc. The conditional is true if any variables were assigned, i.e., if
819 The C</g> modifier specifies global pattern matching--that is,
820 matching as many times as possible within the string. How it behaves
821 depends on the context. In list context, it returns a list of the
822 substrings matched by any capturing parentheses in the regular
823 expression. If there are no parentheses, it returns a list of all
824 the matched strings, as if there were parentheses around the whole
827 In scalar context, each execution of C<m//g> finds the next match,
828 returning true if it matches, and false if there is no further match.
829 The position after the last match can be read or set using the pos()
830 function; see L<perlfunc/pos>. A failed match normally resets the
831 search position to the beginning of the string, but you can avoid that
832 by adding the C</c> modifier (e.g. C<m//gc>). Modifying the target
833 string also resets the search position.
835 You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
836 zero-width assertion that matches the exact position where the previous
837 C<m//g>, if any, left off. The C<\G> assertion is not supported without
838 the C</g> modifier. (Currently, without C</g>, C<\G> behaves just like
839 C<\A>, but that's accidental and may change in the future.)
844 ($one,$five,$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
847 $/ = ""; $* = 1; # $* deprecated in modern perls
848 while (defined($paragraph = <>)) {
849 while ($paragraph =~ /[a-z]['")]*[.!?]+['")]*\s/g) {
853 print "$sentences\n";
855 # using m//gc with \G
859 print $1 while /(o)/gc; print "', pos=", pos, "\n";
861 print $1 if /\G(q)/gc; print "', pos=", pos, "\n";
863 print $1 while /(p)/gc; print "', pos=", pos, "\n";
866 The last example should print:
875 A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
876 combine several regexps like this to process a string part-by-part,
877 doing different actions depending on which regexp matched. Each
878 regexp tries to match where the previous one leaves off.
881 $url = new URI::URL "http://www/"; die if $url eq "xXx";
885 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
886 print(" lowercase"), redo LOOP if /\G[a-z]+\b[,.;]?\s*/gc;
887 print(" UPPERCASE"), redo LOOP if /\G[A-Z]+\b[,.;]?\s*/gc;
888 print(" Capitalized"), redo LOOP if /\G[A-Z][a-z]+\b[,.;]?\s*/gc;
889 print(" MiXeD"), redo LOOP if /\G[A-Za-z]+\b[,.;]?\s*/gc;
890 print(" alphanumeric"), redo LOOP if /\G[A-Za-z0-9]+\b[,.;]?\s*/gc;
891 print(" line-noise"), redo LOOP if /\G[^A-Za-z0-9]+/gc;
892 print ". That's all!\n";
895 Here is the output (split into several lines):
897 line-noise lowercase line-noise lowercase UPPERCASE line-noise
898 UPPERCASE line-noise lowercase line-noise lowercase line-noise
899 lowercase lowercase line-noise lowercase lowercase line-noise
900 MiXeD line-noise. That's all!
906 A single-quoted, literal string. A backslash represents a backslash
907 unless followed by the delimiter or another backslash, in which case
908 the delimiter or backslash is interpolated.
910 $foo = q!I said, "You said, 'She said it.'"!;
911 $bar = q('This is it.');
912 $baz = '\n'; # a two-character string
918 A double-quoted, interpolated string.
921 (*** The previous line contains the naughty word "$1".\n)
922 if /\b(tcl|java|python)\b/i; # :-)
923 $baz = "\n"; # a one-character string
925 =item qr/STRING/imosx
927 This operators quotes--and compiles--its I<STRING> as a regular
928 expression. I<STRING> is interpolated the same way as I<PATTERN>
929 in C<m/PATTERN/>. If "'" is used as the delimiter, no interpolation
930 is done. Returns a Perl value which may be used instead of the
931 corresponding C</STRING/imosx> expression.
935 $rex = qr/my.STRING/is;
942 The result may be used as a subpattern in a match:
945 $string =~ /foo${re}bar/; # can be interpolated in other patterns
946 $string =~ $re; # or used standalone
947 $string =~ /$re/; # or this way
949 Since Perl may compile the pattern at the moment of execution of qr()
950 operator, using qr() may have speed advantages in some situations,
951 notably if the result of qr() is used standalone:
954 my $patterns = shift;
955 my @compiled = map qr/$_/i, @$patterns;
958 foreach my $pat @compiled {
959 $success = 1, last if /$pat/;
965 Precompilation of the pattern into an internal representation at
966 the moment of qr() avoids a need to recompile the pattern every
967 time a match C</$pat/> is attempted. (Perl has many other internal
968 optimizations, but none would be triggered in the above example if
969 we did not use qr() operator.)
973 i Do case-insensitive pattern matching.
974 m Treat string as multiple lines.
975 o Compile pattern only once.
976 s Treat string as single line.
977 x Use extended regular expressions.
979 See L<perlre> for additional information on valid syntax for STRING, and
980 for a detailed look at the semantics of regular expressions.
986 A string which is (possibly) interpolated and then executed as a system
987 command with C</bin/sh> or its equivalent. Shell wildcards, pipes,
988 and redirections will be honored. The collected standard output of the
989 command is returned; standard error is unaffected. In scalar context,
990 it comes back as a single (potentially multi-line) string. In list
991 context, returns a list of lines (however you've defined lines with $/
992 or $INPUT_RECORD_SEPARATOR).
994 Because backticks do not affect standard error, use shell file descriptor
995 syntax (assuming the shell supports this) if you care to address this.
996 To capture a command's STDERR and STDOUT together:
998 $output = `cmd 2>&1`;
1000 To capture a command's STDOUT but discard its STDERR:
1002 $output = `cmd 2>/dev/null`;
1004 To capture a command's STDERR but discard its STDOUT (ordering is
1007 $output = `cmd 2>&1 1>/dev/null`;
1009 To exchange a command's STDOUT and STDERR in order to capture the STDERR
1010 but leave its STDOUT to come out the old STDERR:
1012 $output = `cmd 3>&1 1>&2 2>&3 3>&-`;
1014 To read both a command's STDOUT and its STDERR separately, it's easiest
1015 and safest to redirect them separately to files, and then read from those
1016 files when the program is done:
1018 system("program args 1>/tmp/program.stdout 2>/tmp/program.stderr");
1020 Using single-quote as a delimiter protects the command from Perl's
1021 double-quote interpolation, passing it on to the shell instead:
1023 $perl_info = qx(ps $$); # that's Perl's $$
1024 $shell_info = qx'ps $$'; # that's the new shell's $$
1026 How that string gets evaluated is entirely subject to the command
1027 interpreter on your system. On most platforms, you will have to protect
1028 shell metacharacters if you want them treated literally. This is in
1029 practice difficult to do, as it's unclear how to escape which characters.
1030 See L<perlsec> for a clean and safe example of a manual fork() and exec()
1031 to emulate backticks safely.
1033 On some platforms (notably DOS-like ones), the shell may not be
1034 capable of dealing with multiline commands, so putting newlines in
1035 the string may not get you what you want. You may be able to evaluate
1036 multiple commands in a single line by separating them with the command
1037 separator character, if your shell supports that (e.g. C<;> on many Unix
1038 shells; C<&> on the Windows NT C<cmd> shell).
1040 Beware that some command shells may place restrictions on the length
1041 of the command line. You must ensure your strings don't exceed this
1042 limit after any necessary interpolations. See the platform-specific
1043 release notes for more details about your particular environment.
1045 Using this operator can lead to programs that are difficult to port,
1046 because the shell commands called vary between systems, and may in
1047 fact not be present at all. As one example, the C<type> command under
1048 the POSIX shell is very different from the C<type> command under DOS.
1049 That doesn't mean you should go out of your way to avoid backticks
1050 when they're the right way to get something done. Perl was made to be
1051 a glue language, and one of the things it glues together is commands.
1052 Just understand what you're getting yourself into.
1054 See L<"I/O Operators"> for more discussion.
1058 Evaluates to a list of the words extracted out of STRING, using embedded
1059 whitespace as the word delimiters. It can be understood as being roughly
1062 split(' ', q/STRING/);
1064 the difference being that it generates a real list at compile time. So
1069 is exactly equivalent to the list:
1071 ('foo', 'bar', 'baz')
1073 Some frequently seen examples:
1075 use POSIX qw( setlocale localeconv )
1076 @EXPORT = qw( foo bar baz );
1078 A common mistake is to try to separate the words with comma or to
1079 put comments into a multi-line C<qw>-string. For this reason, the
1080 B<-w> switch (that is, the C<$^W> variable) produces warnings if
1081 the STRING contains the "," or the "#" character.
1083 =item s/PATTERN/REPLACEMENT/egimosx
1085 Searches a string for a pattern, and if found, replaces that pattern
1086 with the replacement text and returns the number of substitutions
1087 made. Otherwise it returns false (specifically, the empty string).
1089 If no string is specified via the C<=~> or C<!~> operator, the C<$_>
1090 variable is searched and modified. (The string specified with C<=~> must
1091 be scalar variable, an array element, a hash element, or an assignment
1092 to one of those, i.e., an lvalue.)
1094 If the delimiter chosen is a single quote, no interpolation is
1095 done on either the PATTERN or the REPLACEMENT. Otherwise, if the
1096 PATTERN contains a $ that looks like a variable rather than an
1097 end-of-string test, the variable will be interpolated into the pattern
1098 at run-time. If you want the pattern compiled only once the first time
1099 the variable is interpolated, use the C</o> option. If the pattern
1100 evaluates to the empty string, the last successfully executed regular
1101 expression is used instead. See L<perlre> for further explanation on these.
1102 See L<perllocale> for discussion of additional considerations that apply
1103 when C<use locale> is in effect.
1107 e Evaluate the right side as an expression.
1108 g Replace globally, i.e., all occurrences.
1109 i Do case-insensitive pattern matching.
1110 m Treat string as multiple lines.
1111 o Compile pattern only once.
1112 s Treat string as single line.
1113 x Use extended regular expressions.
1115 Any non-alphanumeric, non-whitespace delimiter may replace the
1116 slashes. If single quotes are used, no interpretation is done on the
1117 replacement string (the C</e> modifier overrides this, however). Unlike
1118 Perl 4, Perl 5 treats backticks as normal delimiters; the replacement
1119 text is not evaluated as a command. If the
1120 PATTERN is delimited by bracketing quotes, the REPLACEMENT has its own
1121 pair of quotes, which may or may not be bracketing quotes, e.g.,
1122 C<s(foo)(bar)> or C<sE<lt>fooE<gt>/bar/>. A C</e> will cause the
1123 replacement portion to be interpreted as a full-fledged Perl expression
1124 and eval()ed right then and there. It is, however, syntax checked at
1129 s/\bgreen\b/mauve/g; # don't change wintergreen
1131 $path =~ s|/usr/bin|/usr/local/bin|;
1133 s/Login: $foo/Login: $bar/; # run-time pattern
1135 ($foo = $bar) =~ s/this/that/; # copy first, then change
1137 $count = ($paragraph =~ s/Mister\b/Mr./g); # get change-count
1140 s/\d+/$&*2/e; # yields 'abc246xyz'
1141 s/\d+/sprintf("%5d",$&)/e; # yields 'abc 246xyz'
1142 s/\w/$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
1144 s/%(.)/$percent{$1}/g; # change percent escapes; no /e
1145 s/%(.)/$percent{$1} || $&/ge; # expr now, so /e
1146 s/^=(\w+)/&pod($1)/ge; # use function call
1148 # expand variables in $_, but dynamics only, using
1149 # symbolic dereferencing
1152 # /e's can even nest; this will expand
1153 # any embedded scalar variable (including lexicals) in $_
1156 # Delete (most) C comments.
1158 /\* # Match the opening delimiter.
1159 .*? # Match a minimal number of characters.
1160 \*/ # Match the closing delimiter.
1163 s/^\s*(.*?)\s*$/$1/; # trim white space in $_, expensively
1165 for ($variable) { # trim white space in $variable, cheap
1170 s/([^ ]*) *([^ ]*)/$2 $1/; # reverse 1st two fields
1172 Note the use of $ instead of \ in the last example. Unlike
1173 B<sed>, we use the \E<lt>I<digit>E<gt> form in only the left hand side.
1174 Anywhere else it's $E<lt>I<digit>E<gt>.
1176 Occasionally, you can't use just a C</g> to get all the changes
1177 to occur that you might want. Here are two common cases:
1179 # put commas in the right places in an integer
1180 1 while s/(\d)(\d\d\d)(?!\d)/$1,$2/g;
1182 # expand tabs to 8-column spacing
1183 1 while s/\t+/' ' x (length($&)*8 - length($`)%8)/e;
1185 =item tr/SEARCHLIST/REPLACEMENTLIST/cdsUC
1187 =item y/SEARCHLIST/REPLACEMENTLIST/cdsUC
1189 Transliterates all occurrences of the characters found in the search list
1190 with the corresponding character in the replacement list. It returns
1191 the number of characters replaced or deleted. If no string is
1192 specified via the =~ or !~ operator, the $_ string is transliterated. (The
1193 string specified with =~ must be a scalar variable, an array element, a
1194 hash element, or an assignment to one of those, i.e., an lvalue.)
1196 A character range may be specified with a hyphen, so C<tr/A-J/0-9/>
1197 does the same replacement as C<tr/ACEGIBDFHJ/0246813579/>.
1198 For B<sed> devotees, C<y> is provided as a synonym for C<tr>. If the
1199 SEARCHLIST is delimited by bracketing quotes, the REPLACEMENTLIST has
1200 its own pair of quotes, which may or may not be bracketing quotes,
1201 e.g., C<tr[A-Z][a-z]> or C<tr(+\-*/)/ABCD/>.
1203 Note also that the whole range idea is rather unportable between
1204 character sets--and even within character sets they may cause results
1205 you probably didn't expect. A sound principle is to use only ranges
1206 that begin from and end at either alphabets of equal case (a-e, A-E),
1207 or digits (0-4). Anything else is unsafe. If in doubt, spell out the
1208 character sets in full.
1212 c Complement the SEARCHLIST.
1213 d Delete found but unreplaced characters.
1214 s Squash duplicate replaced characters.
1215 U Translate to/from UTF-8.
1216 C Translate to/from 8-bit char (octet).
1218 If the C</c> modifier is specified, the SEARCHLIST character set
1219 is complemented. If the C</d> modifier is specified, any characters
1220 specified by SEARCHLIST not found in REPLACEMENTLIST are deleted.
1221 (Note that this is slightly more flexible than the behavior of some
1222 B<tr> programs, which delete anything they find in the SEARCHLIST,
1223 period.) If the C</s> modifier is specified, sequences of characters
1224 that were transliterated to the same character are squashed down
1225 to a single instance of the character.
1227 If the C</d> modifier is used, the REPLACEMENTLIST is always interpreted
1228 exactly as specified. Otherwise, if the REPLACEMENTLIST is shorter
1229 than the SEARCHLIST, the final character is replicated till it is long
1230 enough. If the REPLACEMENTLIST is empty, the SEARCHLIST is replicated.
1231 This latter is useful for counting characters in a class or for
1232 squashing character sequences in a class.
1234 The first C</U> or C</C> modifier applies to the left side of the translation.
1235 The second one applies to the right side. If present, these modifiers override
1236 the current utf8 state.
1240 $ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case
1242 $cnt = tr/*/*/; # count the stars in $_
1244 $cnt = $sky =~ tr/*/*/; # count the stars in $sky
1246 $cnt = tr/0-9//; # count the digits in $_
1248 tr/a-zA-Z//s; # bookkeeper -> bokeper
1250 ($HOST = $host) =~ tr/a-z/A-Z/;
1252 tr/a-zA-Z/ /cs; # change non-alphas to single space
1255 [\000-\177]; # delete 8th bit
1257 tr/\0-\xFF//CU; # change Latin-1 to Unicode
1258 tr/\0-\x{FF}//UC; # change Unicode to Latin-1
1260 If multiple transliterations are given for a character, only the
1265 will transliterate any A to X.
1267 Because the transliteration table is built at compile time, neither
1268 the SEARCHLIST nor the REPLACEMENTLIST are subjected to double quote
1269 interpolation. That means that if you want to use variables, you
1272 eval "tr/$oldlist/$newlist/";
1275 eval "tr/$oldlist/$newlist/, 1" or die $@;
1279 =head2 Gory details of parsing quoted constructs
1281 When presented with something that might have several different
1282 interpretations, Perl uses the B<DWIM> (that's "Do What I Mean")
1283 principle to pick the most probable interpretation. This strategy
1284 is so successful that Perl programmers often do not suspect the
1285 ambivalence of what they write. But from time to time, Perl's
1286 notions differ substantially from what the author honestly meant.
1288 This section hopes to clarify how Perl handles quoted constructs.
1289 Although the most common reason to learn this is to unravel labyrinthine
1290 regular expressions, because the initial steps of parsing are the
1291 same for all quoting operators, they are all discussed together.
1293 The most important Perl parsing rule is the first one discussed
1294 below: when processing a quoted construct, Perl first finds the end
1295 of that construct, then interprets its contents. If you understand
1296 this rule, you may skip the rest of this section on the first
1297 reading. The other rules are likely to contradict the user's
1298 expectations much less frequently than this first one.
1300 Some passes discussed below are performed concurrently, but because
1301 their results are the same, we consider them individually. For different
1302 quoting constructs, Perl performs different numbers of passes, from
1303 one to five, but these passes are always performed in the same order.
1307 =item Finding the end
1309 The first pass is finding the end of the quoted construct, whether
1310 it be a multicharacter delimiter C<"\nEOF\n"> in the C<<<EOF>
1311 construct, a C</> that terminates a C<qq//> construct, a C<]> which
1312 terminates C<qq[]> construct, or a C<E<gt>> which terminates a
1313 fileglob started with C<E<lt>>.
1315 When searching for single-character non-pairing delimiters, such
1316 as C</>, combinations of C<\\> and C<\/> are skipped. However,
1317 when searching for single-character pairing delimiter like C<[>,
1318 combinations of C<\\>, C<\]>, and C<\[> are all skipped, and nested
1319 C<[>, C<]> are skipped as well. When searching for multicharacter
1320 delimiters, nothing is skipped.
1322 For constructs with three-part delimiters (C<s///>, C<y///>, and
1323 C<tr///>), the search is repeated once more.
1325 During this search no attention is paid to the semantics of the construct.
1328 "$hash{"$foo/$bar"}"
1333 bar # NOT a comment, this slash / terminated m//!
1336 do not form legal quoted expressions. The quoted part ends on the
1337 first C<"> and C</>, and the rest happens to be a syntax error.
1338 Because the slash that terminated C<m//> was followed by a C<SPACE>,
1339 the example above is not C<m//x>, but rather C<m//> with no C</x>
1340 modifier. So the embedded C<#> is interpreted as a literal C<#>.
1342 =item Removal of backslashes before delimiters
1344 During the second pass, text between the starting and ending
1345 delimiters is copied to a safe location, and the C<\> is removed
1346 from combinations consisting of C<\> and delimiter--or delimiters,
1347 meaning both starting and ending delimiters will should these differ.
1348 This removal does not happen for multi-character delimiters.
1349 Note that the combination C<\\> is left intact, just as it was.
1351 Starting from this step no information about the delimiters is
1356 The next step is interpolation in the text obtained, which is now
1357 delimiter-independent. There are four different cases.
1361 =item C<<<'EOF'>, C<m''>, C<s'''>, C<tr///>, C<y///>
1363 No interpolation is performed.
1367 The only interpolation is removal of C<\> from pairs C<\\>.
1369 =item C<"">, C<``>, C<qq//>, C<qx//>, C<<file*globE<gt>>
1371 C<\Q>, C<\U>, C<\u>, C<\L>, C<\l> (possibly paired with C<\E>) are
1372 converted to corresponding Perl constructs. Thus, C<"$foo\Qbaz$bar">
1373 is converted to C<$foo . (quotemeta("baz" . $bar))> internally.
1374 The other combinations are replaced with appropriate expansions.
1376 Let it be stressed that I<whatever falls between C<\Q> and C<\E>>
1377 is interpolated in the usual way. Something like C<"\Q\\E"> has
1378 no C<\E> inside. instead, it has C<\Q>, C<\\>, and C<E>, so the
1379 result is the same as for C<"\\\\E">. As a general rule, backslashes
1380 between C<\Q> and C<\E> may lead to counterintuitive results. So,
1381 C<"\Q\t\E"> is converted to C<quotemeta("\t")>, which is the same
1382 as C<"\\\t"> (since TAB is not alphanumeric). Note also that:
1387 may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
1389 Interpolated scalars and arrays are converted internally to the C<join> and
1390 C<.> catentation operations. Thus, C<"$foo XXX '@arr'"> becomes:
1392 $foo . " XXX '" . (join $", @arr) . "'";
1394 All operations above are performed simultaneously, left to right.
1396 Because the result of C<"\Q STRING \E"> has all metacharacters
1397 quoted, there is no way to insert a literal C<$> or C<@> inside a
1398 C<\Q\E> pair. If protected by C<\>, C<$> will be quoted to became
1399 C<"\\\$">; if not, it is interpreted as the start of an interpolated
1402 Note also that the interpolation code needs to make a decision on
1403 where the interpolated scalar ends. For instance, whether
1404 C<"a $b -E<gt> {c}"> really means:
1406 "a " . $b . " -> {c}";
1412 Most of the time, the longest possible text that does not include
1413 spaces between components and which contains matching braces or
1414 brackets. because the outcome may be determined by voting based
1415 on heuristic estimators, the result is not strictly predictable.
1416 Fortunately, it's usually correct for ambiguous cases.
1418 =item C<?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
1420 Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, and interpolation
1421 happens (almost) as with C<qq//> constructs, but the substitution
1422 of C<\> followed by RE-special chars (including C<\>) is not
1423 performed. Moreover, inside C<(?{BLOCK})>, C<(?# comment )>, and
1424 a C<#>-comment in a C<//x>-regular expression, no processing is
1425 performed whatsoever. This is the first step at which the presence
1426 of the C<//x> modifier is relevant.
1428 Interpolation has several quirks: C<$|>, C<$(>, and C<$)> are not
1429 interpolated, and constructs C<$var[SOMETHING]> are voted (by several
1430 different estimators) to be either an array element or C<$var>
1431 followed by an RE alternative. This is where the notation
1432 C<${arr[$bar]}> comes handy: C</${arr[0-9]}/> is interpreted as
1433 array element C<-9>, not as a regular expression from the variable
1434 C<$arr> followed by a digit, which would be the interpretation of
1435 C</$arr[0-9]/>. Since voting among different estimators may occur,
1436 the result is not predictable.
1438 It is at this step that C<\1> is begrudgingly converted to C<$1> in
1439 the replacement text of C<s///> to correct the incorrigible
1440 I<sed> hackers who haven't picked up the saner idiom yet. A warning
1441 is emitted if the B<-w> command-line flag (that is, the C<$^W> variable)
1444 The lack of processing of C<\\> creates specific restrictions on
1445 the post-processed text. If the delimiter is C</>, one cannot get
1446 the combination C<\/> into the result of this step. C</> will
1447 finish the regular expression, C<\/> will be stripped to C</> on
1448 the previous step, and C<\\/> will be left as is. Because C</> is
1449 equivalent to C<\/> inside a regular expression, this does not
1450 matter unless the delimiter happens to be character special to the
1451 RE engine, such as in C<s*foo*bar*>, C<m[foo]>, or C<?foo?>; or an
1452 alphanumeric char, as in:
1456 In the RE above, which is intentionally obfuscated for illustration, the
1457 delimiter is C<m>, the modifier is C<mx>, and after backslash-removal the
1458 RE is the same as for C<m/ ^ a s* b /mx>). There's more than one
1459 reason you're encouraged to restrict your delimiters to non-alphanumeric,
1460 non-whitespace choices.
1464 This step is the last one for all constructs except regular expressions,
1465 which are processed further.
1467 =item Interpolation of regular expressions
1469 Previous steps were performed during the compilation of Perl code,
1470 but this one happens at run time--although it may be optimized to
1471 be calculated at compile time if appropriate. After preprocessing
1472 described above, and possibly after evaluation if catenation,
1473 joining, casing translation, or metaquoting are involved, the
1474 resulting I<string> is passed to the RE engine for compilation.
1476 Whatever happens in the RE engine might be better discussed in L<perlre>,
1477 but for the sake of continuity, we shall do so here.
1479 This is another step where the presence of the C<//x> modifier is
1480 relevant. The RE engine scans the string from left to right and
1481 converts it to a finite automaton.
1483 Backslashed characters are either replaced with corresponding
1484 literal strings (as with C<\{>), or else they generate special nodes
1485 in the finite automaton (as with C<\b>). Characters special to the
1486 RE engine (such as C<|>) generate corresponding nodes or groups of
1487 nodes. C<(?#...)> comments are ignored. All the rest is either
1488 converted to literal strings to match, or else is ignored (as is
1489 whitespace and C<#>-style comments if C<//x> is present).
1491 Parsing of the bracketed character class construct, C<[...]>, is
1492 rather different than the rule used for the rest of the pattern.
1493 The terminator of this construct is found using the same rules as
1494 for finding the terminator of a C<{}>-delimited construct, the only
1495 exception being that C<]> immediately following C<[> is treated as
1496 though preceded by a backslash. Similarly, the terminator of
1497 C<(?{...})> is found using the same rules as for finding the
1498 terminator of a C<{}>-delimited construct.
1500 It is possible to inspect both the string given to RE engine and the
1501 resulting finite automaton. See the arguments C<debug>/C<debugcolor>
1502 in the C<use L<re>> pragma, as well as Perl's B<-Dr> command-line
1503 switch documented in L<perlrun/Switches>.
1505 =item Optimization of regular expressions
1507 This step is listed for completeness only. Since it does not change
1508 semantics, details of this step are not documented and are subject
1509 to change without notice. This step is performed over the finite
1510 automaton that was generated during the previous pass.
1512 It is at this stage that C<split()> silently optimizes C</^/> to
1517 =head2 I/O Operators
1519 There are several I/O operators you should know about.
1521 A string enclosed by backticks (grave accents) first undergoes
1522 double-quote interpolation. It is then interpreted as an external
1523 command, and the output of that command is the value of the
1525 string consisting of all output is returned. In list context, a
1526 list of values is returned, one per line of output. (You can set
1527 C<$/> to use a different line terminator.) The command is executed
1528 each time the pseudo-literal is evaluated. The status value of the
1529 command is returned in C<$?> (see L<perlvar> for the interpretation
1530 of C<$?>). Unlike in B<csh>, no translation is done on the return
1531 data--newlines remain newlines. Unlike in any of the shells, single
1532 quotes do not hide variable names in the command from interpretation.
1533 To pass a literal dollar-sign through to the shell you need to hide
1534 it with a backslash. The generalized form of backticks is C<qx//>.
1535 (Because backticks always undergo shell expansion as well, see
1536 L<perlsec> for security concerns.)
1538 In scalar context, evaluating a filehandle in angle brackets yields
1539 the next line from that file (the newline, if any, included), or
1540 C<undef> at end-of-file or on error. When C<$/> is set to C<undef>
1541 (sometimes known as file-slurp mode) and the file is empty, it
1542 returns C<''> the first time, followed by C<undef> subsequently.
1544 Ordinarily you must assign the returned value to a variable, but
1545 there is one situation where an automatic assignment happens. If
1546 and only if the input symbol is the only thing inside the conditional
1547 of a C<while> statement (even if disguised as a C<for(;;)> loop),
1548 the value is automatically assigned to the global variable $_,
1549 destroying whatever was there previously. (This may seem like an
1550 odd thing to you, but you'll use the construct in almost every Perl
1551 script you write.) The $_ variables is not implicitly localized.
1552 You'll have to put a C<local $_;> before the loop if you want that
1555 The following lines are equivalent:
1557 while (defined($_ = <STDIN>)) { print; }
1558 while ($_ = <STDIN>) { print; }
1559 while (<STDIN>) { print; }
1560 for (;<STDIN>;) { print; }
1561 print while defined($_ = <STDIN>);
1562 print while ($_ = <STDIN>);
1563 print while <STDIN>;
1565 This also behaves similarly, but avoids $_ :
1567 while (my $line = <STDIN>) { print $line }
1569 In these loop constructs, the assigned value (whether assignment
1570 is automatic or explicit) is then tested to see whether it is
1571 defined. The defined test avoids problems where line has a string
1572 value that would be treated as false by Perl, for example a "" or
1573 a "0" with no trailing newline. If you really mean for such values
1574 to terminate the loop, they should be tested for explicitly:
1576 while (($_ = <STDIN>) ne '0') { ... }
1577 while (<STDIN>) { last unless $_; ... }
1579 In other boolean contexts, C<E<lt>I<filehandle>E<gt>> without an
1580 explicit C<defined> test or comparison elicit a warning if the B<-w>
1581 command-line switch (the C<$^W> variable) is in effect.
1583 The filehandles STDIN, STDOUT, and STDERR are predefined. (The
1584 filehandles C<stdin>, C<stdout>, and C<stderr> will also work except
1585 in packages, where they would be interpreted as local identifiers
1586 rather than global.) Additional filehandles may be created with
1587 the open() function, amongst others. See L<perlopentut> and
1588 L<perlfunc/open> for details on this.
1590 If a E<lt>FILEHANDLEE<gt> is used in a context that is looking for
1591 a list, a list comprising all input lines is returned, one line per
1592 list element. It's easy to grow to a rather large data space this
1593 way, so use with care.
1595 E<lt>FILEHANDLEE<gt> may also be spelled C<readline(*FILEHANDLE)>.
1596 See L<perlfunc/readline>.
1598 The null filehandle E<lt>E<gt> is special: it can be used to emulate the
1599 behavior of B<sed> and B<awk>. Input from E<lt>E<gt> comes either from
1600 standard input, or from each file listed on the command line. Here's
1601 how it works: the first time E<lt>E<gt> is evaluated, the @ARGV array is
1602 checked, and if it is empty, C<$ARGV[0]> is set to "-", which when opened
1603 gives you standard input. The @ARGV array is then processed as a list
1604 of filenames. The loop
1607 ... # code for each line
1610 is equivalent to the following Perl-like pseudo code:
1612 unshift(@ARGV, '-') unless @ARGV;
1613 while ($ARGV = shift) {
1616 ... # code for each line
1620 except that it isn't so cumbersome to say, and will actually work.
1621 It really does shift the @ARGV array and put the current filename
1622 into the $ARGV variable. It also uses filehandle I<ARGV>
1623 internally--E<lt>E<gt> is just a synonym for E<lt>ARGVE<gt>, which
1624 is magical. (The pseudo code above doesn't work because it treats
1625 E<lt>ARGVE<gt> as non-magical.)
1627 You can modify @ARGV before the first E<lt>E<gt> as long as the array ends up
1628 containing the list of filenames you really want. Line numbers (C<$.>)
1629 continue as though the input were one big happy file. See the example
1630 in L<perlfunc/eof> for how to reset line numbers on each file.
1632 If you want to set @ARGV to your own list of files, go right ahead.
1633 This sets @ARGV to all plain text files if no @ARGV was given:
1635 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
1637 You can even set them to pipe commands. For example, this automatically
1638 filters compressed arguments through B<gzip>:
1640 @ARGV = map { /\.(gz|Z)$/ ? "gzip -dc < $_ |" : $_ } @ARGV;
1642 If you want to pass switches into your script, you can use one of the
1643 Getopts modules or put a loop on the front like this:
1645 while ($_ = $ARGV[0], /^-/) {
1648 if (/^-D(.*)/) { $debug = $1 }
1649 if (/^-v/) { $verbose++ }
1650 # ... # other switches
1654 # ... # code for each line
1657 The E<lt>E<gt> symbol will return C<undef> for end-of-file only once.
1658 If you call it again after this, it will assume you are processing another
1659 @ARGV list, and if you haven't set @ARGV, will read input from STDIN.
1661 If angle brackets contain is a simple scalar variable (e.g.,
1662 E<lt>$fooE<gt>), then that variable contains the name of the
1663 filehandle to input from, or its typeglob, or a reference to the
1669 If what's within the angle brackets is neither a filehandle nor a simple
1670 scalar variable containing a filehandle name, typeglob, or typeglob
1671 reference, it is interpreted as a filename pattern to be globbed, and
1672 either a list of filenames or the next filename in the list is returned,
1673 depending on context. This distinction is determined on syntactic
1674 grounds alone. That means C<E<lt>$xE<gt>> is always a readline() from
1675 an indirect handle, but C<E<lt>$hash{key}E<gt>> is always a glob().
1676 That's because $x is a simple scalar variable, but C<$hash{key}> is
1677 not--it's a hash element.
1679 One level of double-quote interpretation is done first, but you can't
1680 say C<E<lt>$fooE<gt>> because that's an indirect filehandle as explained
1681 in the previous paragraph. (In older versions of Perl, programmers
1682 would insert curly brackets to force interpretation as a filename glob:
1683 C<E<lt>${foo}E<gt>>. These days, it's considered cleaner to call the
1684 internal function directly as C<glob($foo)>, which is probably the right
1685 way to have done it in the first place.) For example:
1693 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
1699 In fact, it's currently implemented that way, but this is expected
1700 to be made completely internal in the near future. (Which means
1701 it will not work on filenames with spaces in them unless you have
1702 csh(1) on your machine.) Of course, the shortest way to do the
1707 Because globbing currently invokes a shell, it's often faster to
1708 call readdir() yourself and do your own grep() on the filenames.
1709 Furthermore, due to its current implementation of using a shell,
1710 the glob() routine may get "Arg list too long" errors (unless you've
1711 installed tcsh(1L) as F</bin/csh> or hacked your F<config.sh>).
1713 A (file)glob evaluates its (embedded) argument only when it is
1714 starting a new list. All values must be read before it will start
1715 over. In list context, this isn't important because you automatically
1716 get them all anyway. However, in scalar context the operator returns
1717 the next value each time it's called, or C
1718 run out. As with filehandle reads, an automatic C<defined> is
1719 generated when the glob occurs in the test part of a C<while>,
1720 because legal glob returns (e.g. a file called F<0>) would otherwise
1721 terminate the loop. Again, C<undef> is returned only once. So if
1722 you're expecting a single value from a glob, it is much better to
1725 ($file) = <blurch*>;
1731 because the latter will alternate between returning a filename and
1734 It you're trying to do variable interpolation, it's definitely better
1735 to use the glob() function, because the older notation can cause people
1736 to become confused with the indirect filehandle notation.
1738 @files = glob("$dir/*.[ch]");
1739 @files = glob($files[$i]);
1741 =head2 Constant Folding
1743 Like C, Perl does a certain amount of expression evaluation at
1744 compile time whenever it determines that all arguments to an
1745 operator are static and have no side effects. In particular, string
1746 concatenation happens at compile time between literals that don't do
1747 variable substitution. Backslash interpolation also happens at
1748 compile time. You can say
1750 'Now is the time for all' . "\n" .
1751 'good men to come to.'
1753 and this all reduces to one string internally. Likewise, if
1756 foreach $file (@filenames) {
1757 if (-s $file > 5 + 100 * 2**16) { }
1760 the compiler will precompute the number which that expression
1761 represents so that the interpreter won't have to.
1763 =head2 Bitwise String Operators
1765 Bitstrings of any size may be manipulated by the bitwise operators
1768 If the operands to a binary bitwise op are strings of different
1769 sizes, B<|> and B<^> ops act as though the shorter operand had
1770 additional zero bits on the right, while the B<&> op acts as though
1771 the longer operand were truncated to the length of the shorter.
1772 The granularity for such extension or truncation is one or more
1775 # ASCII-based examples
1776 print "j p \n" ^ " a h"; # prints "JAPH\n"
1777 print "JA" | " ph\n"; # prints "japh\n"
1778 print "japh\nJunk" & '_____'; # prints "JAPH\n";
1779 print 'p N$' ^ " E<H\n"; # prints "Perl\n";
1781 If you are intending to manipulate bitstrings, be certain that
1782 you're supplying bitstrings: If an operand is a number, that will imply
1783 a B<numeric> bitwise operation. You may explicitly show which type of
1784 operation you intend by using C<""> or C<0+>, as in the examples below.
1786 $foo = 150 | 105 ; # yields 255 (0x96 | 0x69 is 0xFF)
1787 $foo = '150' | 105 ; # yields 255
1788 $foo = 150 | '105'; # yields 255
1789 $foo = '150' | '105'; # yields string '155' (under ASCII)
1791 $baz = 0+$foo & 0+$bar; # both ops explicitly numeric
1792 $biz = "$foo" ^ "$bar"; # both ops explicitly stringy
1794 See L<perlfunc/vec> for information on how to manipulate individual bits
1797 =head2 Integer Arithmetic
1799 By default, Perl assumes that it must do most of its arithmetic in
1800 floating point. But by saying
1804 you may tell the compiler that it's okay to use integer operations
1805 (if it feels like it) from here to the end of the enclosing BLOCK.
1806 An inner BLOCK may countermand this by saying
1810 which lasts until the end of that BLOCK. Note that this doesn't
1811 mean everything is only an integer, merely that Perl may use integer
1812 operations if it is so inclined. For example, even under C<use
1813 integer>, if you take the C<sqrt(2)>, you'll still get C<1.4142135623731>
1816 Used on numbers, the bitwise operators ("&", "|", "^", "~", "<<",
1817 and ">>") always produce integral results. (But see also L<Bitwise
1818 String Operators>.) However, C<use integer> still has meaning for
1819 them. By default, their results are interpreted as unsigned integers, but
1820 if C<use integer> is in effect, their results are interpreted
1821 as signed integers. For example, C<~0> usually evaluates to a large
1822 integral value. However, C<use integer; ~0> is C<-1> on twos-complement
1825 =head2 Floating-point Arithmetic
1827 While C<use integer> provides integer-only arithmetic, there is no
1828 analogous mechanism to provide automatic rounding or truncation to a
1829 certain number of decimal places. For rounding to a certain number
1830 of digits, sprintf() or printf() is usually the easiest route.
1833 Floating-point numbers are only approximations to what a mathematician
1834 would call real numbers. There are infinitely more reals than floats,
1835 so some corners must be cut. For example:
1837 printf "%.20g\n", 123456789123456789;
1838 # produces 123456789123456784
1840 Testing for exact equality of floating-point equality or inequality is
1841 not a good idea. Here's a (relatively expensive) work-around to compare
1842 whether two floating-point numbers are equal to a particular number of
1843 decimal places. See Knuth, volume II, for a more robust treatment of
1847 my ($X, $Y, $POINTS) = @_;
1849 $tX = sprintf("%.${POINTS}g", $X);
1850 $tY = sprintf("%.${POINTS}g", $Y);
1854 The POSIX module (part of the standard perl distribution) implements
1855 ceil(), floor(), and other mathematical and trigonometric functions.
1856 The Math::Complex module (part of the standard perl distribution)
1857 defines mathematical functions that work on both the reals and the
1858 imaginary numbers. Math::Complex not as efficient as POSIX, but
1859 POSIX can't work with complex numbers.
1861 Rounding in financial applications can have serious implications, and
1862 the rounding method used should be specified precisely. In these
1863 cases, it probably pays not to trust whichever system rounding is
1864 being used by Perl, but to instead implement the rounding function you
1867 =head2 Bigger Numbers
1869 The standard Math::BigInt and Math::BigFloat modules provide
1870 variable-precision arithmetic and overloaded operators, although
1871 they're currently pretty slow. At the cost of some space and
1872 considerable speed, they avoid the normal pitfalls associated with
1873 limited-precision representations.
1876 $x = Math::BigInt->new('123456789123456789');
1879 # prints +15241578780673678515622620750190521
1881 The non-standard modules SSLeay::BN and Math::Pari provide
1882 equivalent functionality (and much more) with a substantial
1883 performance savings.