3 perlop - Perl operators and precedence
7 =head2 Operator Precedence and Associativity
9 Operator precedence and associativity work in Perl more or less like
10 they do in mathematics.
12 I<Operator precedence> means some operators are evaluated before
13 others. For example, in C<2 + 4 * 5>, the multiplication has higher
14 precedence so C<4 * 5> is evaluated first yielding C<2 + 20 ==
15 22> and not C<6 * 5 == 30>.
17 I<Operator associativity> defines what happens if a sequence of the
18 same operators is used one after another: whether the evaluator will
19 evaluate the left operations first or the right. For example, in C<8
20 - 4 - 2>, subtraction is left associative so Perl evaluates the
21 expression left to right. C<8 - 4> is evaluated first making the
22 expression C<4 - 2 == 2> and not C<8 - 2 == 6>.
24 Perl operators have the following associativity and precedence,
25 listed from highest precedence to lowest. Operators borrowed from
26 C keep the same precedence relationship with each other, even where
27 C's precedence is slightly screwy. (This makes learning Perl easier
28 for C folks.) With very few exceptions, these all operate on scalar
29 values only, not array values.
31 left terms and list operators (leftward)
35 right ! ~ \ and unary + and -
40 nonassoc named unary operators
41 nonassoc < > <= >= lt gt le ge
42 nonassoc == != <=> eq ne cmp
51 nonassoc list operators (rightward)
56 In the following sections, these operators are covered in precedence order.
58 Many operators can be overloaded for objects. See L<overload>.
60 =head2 Terms and List Operators (Leftward)
62 A TERM has the highest precedence in Perl. They include variables,
63 quote and quote-like operators, any expression in parentheses,
64 and any function whose arguments are parenthesized. Actually, there
65 aren't really functions in this sense, just list operators and unary
66 operators behaving as functions because you put parentheses around
67 the arguments. These are all documented in L<perlfunc>.
69 If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
70 is followed by a left parenthesis as the next token, the operator and
71 arguments within parentheses are taken to be of highest precedence,
72 just like a normal function call.
74 In the absence of parentheses, the precedence of list operators such as
75 C<print>, C<sort>, or C<chmod> is either very high or very low depending on
76 whether you are looking at the left side or the right side of the operator.
79 @ary = (1, 3, sort 4, 2);
80 print @ary; # prints 1324
82 the commas on the right of the sort are evaluated before the sort,
83 but the commas on the left are evaluated after. In other words,
84 list operators tend to gobble up all arguments that follow, and
85 then act like a simple TERM with regard to the preceding expression.
86 Be careful with parentheses:
88 # These evaluate exit before doing the print:
89 print($foo, exit); # Obviously not what you want.
90 print $foo, exit; # Nor is this.
92 # These do the print before evaluating exit:
93 (print $foo), exit; # This is what you want.
94 print($foo), exit; # Or this.
95 print ($foo), exit; # Or even this.
99 print ($foo & 255) + 1, "\n";
101 probably doesn't do what you expect at first glance. The parentheses
102 enclose the argument list for C<print> which is evaluated (printing
103 the result of C<$foo & 255>). Then one is added to the return value
104 of C<print> (usually 1). The result is something like this:
106 1 + 1, "\n"; # Obviously not what you meant.
108 To do what you meant properly, you must write:
110 print(($foo & 255) + 1, "\n");
112 See L<Named Unary Operators> for more discussion of this.
114 Also parsed as terms are the C<do {}> and C<eval {}> constructs, as
115 well as subroutine and method calls, and the anonymous
116 constructors C<[]> and C<{}>.
118 See also L<Quote and Quote-like Operators> toward the end of this section,
119 as well as L<"I/O Operators">.
121 =head2 The Arrow Operator
123 "C<< -> >>" is an infix dereference operator, just as it is in C
124 and C++. If the right side is either a C<[...]>, C<{...}>, or a
125 C<(...)> subscript, then the left side must be either a hard or
126 symbolic reference to an array, a hash, or a subroutine respectively.
127 (Or technically speaking, a location capable of holding a hard
128 reference, if it's an array or hash reference being used for
129 assignment.) See L<perlreftut> and L<perlref>.
131 Otherwise, the right side is a method name or a simple scalar
132 variable containing either the method name or a subroutine reference,
133 and the left side must be either an object (a blessed reference)
134 or a class name (that is, a package name). See L<perlobj>.
136 =head2 Auto-increment and Auto-decrement
138 "++" and "--" work as in C. That is, if placed before a variable,
139 they increment or decrement the variable by one before returning the
140 value, and if placed after, increment or decrement after returning the
144 print $i++; # prints 0
145 print ++$j; # prints 1
147 The auto-increment operator has a little extra builtin magic to it. If
148 you increment a variable that is numeric, or that has ever been used in
149 a numeric context, you get a normal increment. If, however, the
150 variable has been used in only string contexts since it was set, and
151 has a value that is not the empty string and matches the pattern
152 C</^[a-zA-Z]*[0-9]*\z/>, the increment is done as a string, preserving each
153 character within its range, with carry:
155 print ++($foo = '99'); # prints '100'
156 print ++($foo = 'a0'); # prints 'a1'
157 print ++($foo = 'Az'); # prints 'Ba'
158 print ++($foo = 'zz'); # prints 'aaa'
160 C<undef> is always treated as numeric, and in particular is changed
161 to C<0> before incrementing (so that a post-increment of an undef value
162 will return C<0> rather than C<undef>).
164 The auto-decrement operator is not magical.
166 =head2 Exponentiation
168 Binary "**" is the exponentiation operator. It binds even more
169 tightly than unary minus, so -2**4 is -(2**4), not (-2)**4. (This is
170 implemented using C's pow(3) function, which actually works on doubles
173 =head2 Symbolic Unary Operators
175 Unary "!" performs logical negation, i.e., "not". See also C<not> for a lower
176 precedence version of this.
178 Unary "-" performs arithmetic negation if the operand is numeric. If
179 the operand is an identifier, a string consisting of a minus sign
180 concatenated with the identifier is returned. Otherwise, if the string
181 starts with a plus or minus, a string starting with the opposite sign
182 is returned. One effect of these rules is that C<-bareword> is equivalent
185 Unary "~" performs bitwise negation, i.e., 1's complement. For
186 example, C<0666 & ~027> is 0640. (See also L<Integer Arithmetic> and
187 L<Bitwise String Operators>.) Note that the width of the result is
188 platform-dependent: ~0 is 32 bits wide on a 32-bit platform, but 64
189 bits wide on a 64-bit platform, so if you are expecting a certain bit
190 width, remember to use the & operator to mask off the excess bits.
192 Unary "+" has no effect whatsoever, even on strings. It is useful
193 syntactically for separating a function name from a parenthesized expression
194 that would otherwise be interpreted as the complete list of function
195 arguments. (See examples above under L<Terms and List Operators (Leftward)>.)
197 Unary "\" creates a reference to whatever follows it. See L<perlreftut>
198 and L<perlref>. Do not confuse this behavior with the behavior of
199 backslash within a string, although both forms do convey the notion
200 of protecting the next thing from interpolation.
202 =head2 Binding Operators
204 Binary "=~" binds a scalar expression to a pattern match. Certain operations
205 search or modify the string $_ by default. This operator makes that kind
206 of operation work on some other string. The right argument is a search
207 pattern, substitution, or transliteration. The left argument is what is
208 supposed to be searched, substituted, or transliterated instead of the default
209 $_. When used in scalar context, the return value generally indicates the
210 success of the operation. Behavior in list context depends on the particular
211 operator. See L</"Regexp Quote-Like Operators"> for details.
213 If the right argument is an expression rather than a search pattern,
214 substitution, or transliteration, it is interpreted as a search pattern at run
217 Binary "!~" is just like "=~" except the return value is negated in
220 =head2 Multiplicative Operators
222 Binary "*" multiplies two numbers.
224 Binary "/" divides two numbers.
226 Binary "%" computes the modulus of two numbers. Given integer
227 operands C<$a> and C<$b>: If C<$b> is positive, then C<$a % $b> is
228 C<$a> minus the largest multiple of C<$b> that is not greater than
229 C<$a>. If C<$b> is negative, then C<$a % $b> is C<$a> minus the
230 smallest multiple of C<$b> that is not less than C<$a> (i.e. the
231 result will be less than or equal to zero).
232 Note that when C<use integer> is in scope, "%" gives you direct access
233 to the modulus operator as implemented by your C compiler. This
234 operator is not as well defined for negative operands, but it will
237 Binary "x" is the repetition operator. In scalar context or if the left
238 operand is not enclosed in parentheses, it returns a string consisting
239 of the left operand repeated the number of times specified by the right
240 operand. In list context, if the left operand is enclosed in
241 parentheses, it repeats the list.
243 print '-' x 80; # print row of dashes
245 print "\t" x ($tab/8), ' ' x ($tab%8); # tab over
247 @ones = (1) x 80; # a list of 80 1's
248 @ones = (5) x @ones; # set all elements to 5
251 =head2 Additive Operators
253 Binary "+" returns the sum of two numbers.
255 Binary "-" returns the difference of two numbers.
257 Binary "." concatenates two strings.
259 =head2 Shift Operators
261 Binary "<<" returns the value of its left argument shifted left by the
262 number of bits specified by the right argument. Arguments should be
263 integers. (See also L<Integer Arithmetic>.)
265 Binary ">>" returns the value of its left argument shifted right by
266 the number of bits specified by the right argument. Arguments should
267 be integers. (See also L<Integer Arithmetic>.)
269 Note that both "<<" and ">>" in Perl are implemented directly using
270 "<<" and ">>" in C. If C<use integer> (see L<Integer Arithmetic>) is
271 in force then signed C integers are used, else unsigned C integers are
272 used. Either way, the implementation isn't going to generate results
273 larger than the size of the integer type Perl was built with (32 bits
276 The result of overflowing the range of the integers is undefined
277 because it is undefined also in C. In other words, using 32-bit
278 integers, C<< 1 << 32 >> is undefined. Shifting by a negative number
279 of bits is also undefined.
281 =head2 Named Unary Operators
283 The various named unary operators are treated as functions with one
284 argument, with optional parentheses. These include the filetest
285 operators, like C<-f>, C<-M>, etc. See L<perlfunc>.
287 If any list operator (print(), etc.) or any unary operator (chdir(), etc.)
288 is followed by a left parenthesis as the next token, the operator and
289 arguments within parentheses are taken to be of highest precedence,
290 just like a normal function call. For example,
291 because named unary operators are higher precedence than ||:
293 chdir $foo || die; # (chdir $foo) || die
294 chdir($foo) || die; # (chdir $foo) || die
295 chdir ($foo) || die; # (chdir $foo) || die
296 chdir +($foo) || die; # (chdir $foo) || die
298 but, because * is higher precedence than named operators:
300 chdir $foo * 20; # chdir ($foo * 20)
301 chdir($foo) * 20; # (chdir $foo) * 20
302 chdir ($foo) * 20; # (chdir $foo) * 20
303 chdir +($foo) * 20; # chdir ($foo * 20)
305 rand 10 * 20; # rand (10 * 20)
306 rand(10) * 20; # (rand 10) * 20
307 rand (10) * 20; # (rand 10) * 20
308 rand +(10) * 20; # rand (10 * 20)
310 See also L<"Terms and List Operators (Leftward)">.
312 =head2 Relational Operators
314 Binary "<" returns true if the left argument is numerically less than
317 Binary ">" returns true if the left argument is numerically greater
318 than the right argument.
320 Binary "<=" returns true if the left argument is numerically less than
321 or equal to the right argument.
323 Binary ">=" returns true if the left argument is numerically greater
324 than or equal to the right argument.
326 Binary "lt" returns true if the left argument is stringwise less than
329 Binary "gt" returns true if the left argument is stringwise greater
330 than the right argument.
332 Binary "le" returns true if the left argument is stringwise less than
333 or equal to the right argument.
335 Binary "ge" returns true if the left argument is stringwise greater
336 than or equal to the right argument.
338 =head2 Equality Operators
340 Binary "==" returns true if the left argument is numerically equal to
343 Binary "!=" returns true if the left argument is numerically not equal
344 to the right argument.
346 Binary "<=>" returns -1, 0, or 1 depending on whether the left
347 argument is numerically less than, equal to, or greater than the right
348 argument. If your platform supports NaNs (not-a-numbers) as numeric
349 values, using them with "<=>" returns undef. NaN is not "<", "==", ">",
350 "<=" or ">=" anything (even NaN), so those 5 return false. NaN != NaN
351 returns true, as does NaN != anything else. If your platform doesn't
352 support NaNs then NaN is just a string with numeric value 0.
354 perl -le '$a = NaN; print "No NaN support here" if $a == $a'
355 perl -le '$a = NaN; print "NaN support here" if $a != $a'
357 Binary "eq" returns true if the left argument is stringwise equal to
360 Binary "ne" returns true if the left argument is stringwise not equal
361 to the right argument.
363 Binary "cmp" returns -1, 0, or 1 depending on whether the left
364 argument is stringwise less than, equal to, or greater than the right
367 "lt", "le", "ge", "gt" and "cmp" use the collation (sort) order specified
368 by the current locale if C<use locale> is in effect. See L<perllocale>.
372 Binary "&" returns its operands ANDed together bit by bit.
373 (See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
375 Note that "&" has lower priority than relational operators, so for example
376 the brackets are essential in a test like
378 print "Even\n" if ($x & 1) == 0;
380 =head2 Bitwise Or and Exclusive Or
382 Binary "|" returns its operands ORed together bit by bit.
383 (See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
385 Binary "^" returns its operands XORed together bit by bit.
386 (See also L<Integer Arithmetic> and L<Bitwise String Operators>.)
388 Note that "|" and "^" have lower priority than relational operators, so
389 for example the brackets are essential in a test like
391 print "false\n" if (8 | 2) != 10;
393 =head2 C-style Logical And
395 Binary "&&" performs a short-circuit logical AND operation. That is,
396 if the left operand is false, the right operand is not even evaluated.
397 Scalar or list context propagates down to the right operand if it
400 =head2 C-style Logical Or
402 Binary "||" performs a short-circuit logical OR operation. That is,
403 if the left operand is true, the right operand is not even evaluated.
404 Scalar or list context propagates down to the right operand if it
407 =head2 C-style Logical Defined-Or
409 Although it has no direct equivalent in C, Perl's C<//> operator is related
410 to its C-style or. In fact, it's exactly the same as C<||>, except that it
411 tests the left hand side's definedness instead of its truth. Thus, C<$a // $b>
412 is similar to C<defined($a) || $b> (except that it returns the value of C<$a>
413 rather than the value of C<defined($a)>) and is exactly equivalent to
414 C<defined($a) ? $a : $b>. This is very useful for providing default values
415 for variables. If you actually want to test if at least one of C<$a> and
416 C<$b> is defined, use C<defined($a // $b)>.
418 The C<||>, C<//> and C<&&> operators return the last value evaluated
419 (unlike C's C<||> and C<&&>, which return 0 or 1). Thus, a reasonably
420 portable way to find out the home directory might be:
422 $home = $ENV{'HOME'} // $ENV{'LOGDIR'} //
423 (getpwuid($<))[7] // die "You're homeless!\n";
425 In particular, this means that you shouldn't use this
426 for selecting between two aggregates for assignment:
428 @a = @b || @c; # this is wrong
429 @a = scalar(@b) || @c; # really meant this
430 @a = @b ? @b : @c; # this works fine, though
432 As more readable alternatives to C<&&>, C<//> and C<||> when used for
433 control flow, Perl provides C<and>, C<err> and C<or> operators (see below).
434 The short-circuit behavior is identical. The precedence of "and", "err"
435 and "or" is much lower, however, so that you can safely use them after a
436 list operator without the need for parentheses:
438 unlink "alpha", "beta", "gamma"
439 or gripe(), next LINE;
441 With the C-style operators that would have been written like this:
443 unlink("alpha", "beta", "gamma")
444 || (gripe(), next LINE);
446 Using "or" for assignment is unlikely to do what you want; see below.
448 =head2 Range Operators
450 Binary ".." is the range operator, which is really two different
451 operators depending on the context. In list context, it returns a
452 list of values counting (up by ones) from the left value to the right
453 value. If the left value is greater than the right value then it
454 returns the empty list. The range operator is useful for writing
455 C<foreach (1..10)> loops and for doing slice operations on arrays. In
456 the current implementation, no temporary array is created when the
457 range operator is used as the expression in C<foreach> loops, but older
458 versions of Perl might burn a lot of memory when you write something
461 for (1 .. 1_000_000) {
465 The range operator also works on strings, using the magical auto-increment,
468 In scalar context, ".." returns a boolean value. The operator is
469 bistable, like a flip-flop, and emulates the line-range (comma) operator
470 of B<sed>, B<awk>, and various editors. Each ".." operator maintains its
471 own boolean state. It is false as long as its left operand is false.
472 Once the left operand is true, the range operator stays true until the
473 right operand is true, I<AFTER> which the range operator becomes false
474 again. It doesn't become false till the next time the range operator is
475 evaluated. It can test the right operand and become false on the same
476 evaluation it became true (as in B<awk>), but it still returns true once.
477 If you don't want it to test the right operand till the next
478 evaluation, as in B<sed>, just use three dots ("...") instead of
479 two. In all other regards, "..." behaves just like ".." does.
481 The right operand is not evaluated while the operator is in the
482 "false" state, and the left operand is not evaluated while the
483 operator is in the "true" state. The precedence is a little lower
484 than || and &&. The value returned is either the empty string for
485 false, or a sequence number (beginning with 1) for true. The
486 sequence number is reset for each range encountered. The final
487 sequence number in a range has the string "E0" appended to it, which
488 doesn't affect its numeric value, but gives you something to search
489 for if you want to exclude the endpoint. You can exclude the
490 beginning point by waiting for the sequence number to be greater
493 If either operand of scalar ".." is a constant expression,
494 that operand is considered true if it is equal (C<==>) to the current
495 input line number (the C<$.> variable).
497 To be pedantic, the comparison is actually C<int(EXPR) == int(EXPR)>,
498 but that is only an issue if you use a floating point expression; when
499 implicitly using C<$.> as described in the previous paragraph, the
500 comparison is C<int(EXPR) == int($.)> which is only an issue when C<$.>
501 is set to a floating point value and you are not reading from a file.
502 Furthermore, C<"span" .. "spat"> or C<2.18 .. 3.14> will not do what
503 you want in scalar context because each of the operands are evaluated
504 using their integer representation.
508 As a scalar operator:
510 if (101 .. 200) { print; } # print 2nd hundred lines, short for
511 # if ($. == 101 .. $. == 200) ...
512 next line if (1 .. /^$/); # skip header lines, short for
513 # ... if ($. == 1 .. /^$/);
514 s/^/> / if (/^$/ .. eof()); # quote body
516 # parse mail messages
518 $in_header = 1 .. /^$/;
519 $in_body = /^$/ .. eof;
526 close ARGV if eof; # reset $. each file
531 for (101 .. 200) { print; } # print $_ 100 times
532 @foo = @foo[0 .. $#foo]; # an expensive no-op
533 @foo = @foo[$#foo-4 .. $#foo]; # slice last 5 items
535 The range operator (in list context) makes use of the magical
536 auto-increment algorithm if the operands are strings. You
539 @alphabet = ('A' .. 'Z');
541 to get all normal letters of the English alphabet, or
543 $hexdigit = (0 .. 9, 'a' .. 'f')[$num & 15];
545 to get a hexadecimal digit, or
547 @z2 = ('01' .. '31'); print $z2[$mday];
549 to get dates with leading zeros. If the final value specified is not
550 in the sequence that the magical increment would produce, the sequence
551 goes until the next value would be longer than the final value
554 Because each operand is evaluated in integer form, C<2.18 .. 3.14> will
555 return two elements in list context.
557 @list = (2.18 .. 3.14); # same as @list = (2 .. 3);
559 =head2 Conditional Operator
561 Ternary "?:" is the conditional operator, just as in C. It works much
562 like an if-then-else. If the argument before the ? is true, the
563 argument before the : is returned, otherwise the argument after the :
564 is returned. For example:
566 printf "I have %d dog%s.\n", $n,
567 ($n == 1) ? '' : "s";
569 Scalar or list context propagates downward into the 2nd
570 or 3rd argument, whichever is selected.
572 $a = $ok ? $b : $c; # get a scalar
573 @a = $ok ? @b : @c; # get an array
574 $a = $ok ? @b : @c; # oops, that's just a count!
576 The operator may be assigned to if both the 2nd and 3rd arguments are
577 legal lvalues (meaning that you can assign to them):
579 ($a_or_b ? $a : $b) = $c;
581 Because this operator produces an assignable result, using assignments
582 without parentheses will get you in trouble. For example, this:
584 $a % 2 ? $a += 10 : $a += 2
588 (($a % 2) ? ($a += 10) : $a) += 2
592 ($a % 2) ? ($a += 10) : ($a += 2)
594 That should probably be written more simply as:
596 $a += ($a % 2) ? 10 : 2;
598 =head2 Assignment Operators
600 "=" is the ordinary assignment operator.
602 Assignment operators work as in C. That is,
610 although without duplicating any side effects that dereferencing the lvalue
611 might trigger, such as from tie(). Other assignment operators work similarly.
612 The following are recognized:
619 Although these are grouped by family, they all have the precedence
622 Unlike in C, the scalar assignment operator produces a valid lvalue.
623 Modifying an assignment is equivalent to doing the assignment and
624 then modifying the variable that was assigned to. This is useful
625 for modifying a copy of something, like this:
627 ($tmp = $global) =~ tr [A-Z] [a-z];
638 Similarly, a list assignment in list context produces the list of
639 lvalues assigned to, and a list assignment in scalar context returns
640 the number of elements produced by the expression on the right hand
641 side of the assignment.
643 =head2 Comma Operator
645 Binary "," is the comma operator. In scalar context it evaluates
646 its left argument, throws that value away, then evaluates its right
647 argument and returns that value. This is just like C's comma operator.
649 In list context, it's just the list argument separator, and inserts
650 both its arguments into the list.
652 The C<< => >> operator is a synonym for the comma, but forces any word
653 to its left to be interpreted as a string (as of 5.001). It is helpful
654 in documenting the correspondence between keys and values in hashes,
655 and other paired elements in lists.
657 =head2 List Operators (Rightward)
659 On the right side of a list operator, it has very low precedence,
660 such that it controls all comma-separated expressions found there.
661 The only operators with lower precedence are the logical operators
662 "and", "or", and "not", which may be used to evaluate calls to list
663 operators without the need for extra parentheses:
665 open HANDLE, "filename"
666 or die "Can't open: $!\n";
668 See also discussion of list operators in L<Terms and List Operators (Leftward)>.
672 Unary "not" returns the logical negation of the expression to its right.
673 It's the equivalent of "!" except for the very low precedence.
677 Binary "and" returns the logical conjunction of the two surrounding
678 expressions. It's equivalent to && except for the very low
679 precedence. This means that it short-circuits: i.e., the right
680 expression is evaluated only if the left expression is true.
682 =head2 Logical or, Defined or, and Exclusive Or
684 Binary "or" returns the logical disjunction of the two surrounding
685 expressions. It's equivalent to || except for the very low precedence.
686 This makes it useful for control flow
688 print FH $data or die "Can't write to FH: $!";
690 This means that it short-circuits: i.e., the right expression is evaluated
691 only if the left expression is false. Due to its precedence, you should
692 probably avoid using this for assignment, only for control flow.
694 $a = $b or $c; # bug: this is wrong
695 ($a = $b) or $c; # really means this
696 $a = $b || $c; # better written this way
698 However, when it's a list-context assignment and you're trying to use
699 "||" for control flow, you probably need "or" so that the assignment
700 takes higher precedence.
702 @info = stat($file) || die; # oops, scalar sense of stat!
703 @info = stat($file) or die; # better, now @info gets its due
705 Then again, you could always use parentheses.
707 Binary "err" is equivalent to C<//>--it's just like binary "or", except it tests
708 its left argument's definedness instead of its truth. There are two ways to
709 remember "err": either because many functions return C<undef> on an B<err>or,
710 or as a sort of correction: C<$a=($b err 'default')>
712 Binary "xor" returns the exclusive-OR of the two surrounding expressions.
713 It cannot short circuit, of course.
715 =head2 C Operators Missing From Perl
717 Here is what C has that Perl doesn't:
723 Address-of operator. (But see the "\" operator for taking a reference.)
727 Dereference-address operator. (Perl's prefix dereferencing
728 operators are typed: $, @, %, and &.)
732 Type-casting operator.
736 =head2 Quote and Quote-like Operators
738 While we usually think of quotes as literal values, in Perl they
739 function as operators, providing various kinds of interpolating and
740 pattern matching capabilities. Perl provides customary quote characters
741 for these behaviors, but also provides a way for you to choose your
742 quote character for any of them. In the following table, a C<{}> represents
743 any pair of delimiters you choose.
745 Customary Generic Meaning Interpolates
750 // m{} Pattern match yes*
752 s{}{} Substitution yes*
753 tr{}{} Transliteration no (but see below)
756 * unless the delimiter is ''.
758 Non-bracketing delimiters use the same character fore and aft, but the four
759 sorts of brackets (round, angle, square, curly) will all nest, which means
768 Note, however, that this does not always work for quoting Perl code:
770 $s = q{ if($a eq "}") ... }; # WRONG
772 is a syntax error. The C<Text::Balanced> module (from CPAN, and
773 starting from Perl 5.8 part of the standard distribution) is able
776 There can be whitespace between the operator and the quoting
777 characters, except when C<#> is being used as the quoting character.
778 C<q#foo#> is parsed as the string C<foo>, while C<q #foo#> is the
779 operator C<q> followed by a comment. Its argument will be taken
780 from the next line. This allows you to write:
782 s {foo} # Replace foo
785 The following escape sequences are available in constructs that interpolate
786 and in transliterations.
793 \a alarm (bell) (BEL)
795 \033 octal char (ESC)
797 \x{263a} wide hex char (SMILEY)
798 \c[ control char (ESC)
799 \N{name} named Unicode character
801 The following escape sequences are available in constructs that interpolate
802 but not in transliterations.
804 \l lowercase next char
805 \u uppercase next char
808 \E end case modification
809 \Q quote non-word characters till \E
811 If C<use locale> is in effect, the case map used by C<\l>, C<\L>,
812 C<\u> and C<\U> is taken from the current locale. See L<perllocale>.
813 If Unicode (for example, C<\N{}> or wide hex characters of 0x100 or
814 beyond) is being used, the case map used by C<\l>, C<\L>, C<\u> and
815 C<\U> is as defined by Unicode. For documentation of C<\N{name}>,
818 All systems use the virtual C<"\n"> to represent a line terminator,
819 called a "newline". There is no such thing as an unvarying, physical
820 newline character. It is only an illusion that the operating system,
821 device drivers, C libraries, and Perl all conspire to preserve. Not all
822 systems read C<"\r"> as ASCII CR and C<"\n"> as ASCII LF. For example,
823 on a Mac, these are reversed, and on systems without line terminator,
824 printing C<"\n"> may emit no actual data. In general, use C<"\n"> when
825 you mean a "newline" for your system, but use the literal ASCII when you
826 need an exact character. For example, most networking protocols expect
827 and prefer a CR+LF (C<"\015\012"> or C<"\cM\cJ">) for line terminators,
828 and although they often accept just C<"\012">, they seldom tolerate just
829 C<"\015">. If you get in the habit of using C<"\n"> for networking,
830 you may be burned some day.
832 For constructs that do interpolate, variables beginning with "C<$>"
833 or "C<@>" are interpolated. Subscripted variables such as C<$a[3]> or
834 C<< $href->{key}[0] >> are also interpolated, as are array and hash slices.
835 But method calls such as C<< $obj->meth >> are not.
837 Interpolating an array or slice interpolates the elements in order,
838 separated by the value of C<$">, so is equivalent to interpolating
839 C<join $", @array>. "Punctuation" arrays such as C<@+> are only
840 interpolated if the name is enclosed in braces C<@{+}>.
842 You cannot include a literal C<$> or C<@> within a C<\Q> sequence.
843 An unescaped C<$> or C<@> interpolates the corresponding variable,
844 while escaping will cause the literal string C<\$> to be inserted.
845 You'll need to write something like C<m/\Quser\E\@\Qhost/>.
847 Patterns are subject to an additional level of interpretation as a
848 regular expression. This is done as a second pass, after variables are
849 interpolated, so that regular expressions may be incorporated into the
850 pattern from the variables. If this is not what you want, use C<\Q> to
851 interpolate a variable literally.
853 Apart from the behavior described above, Perl does not expand
854 multiple levels of interpolation. In particular, contrary to the
855 expectations of shell programmers, back-quotes do I<NOT> interpolate
856 within double quotes, nor do single quotes impede evaluation of
857 variables when used within double quotes.
859 =head2 Regexp Quote-Like Operators
861 Here are the quote-like operators that apply to pattern
862 matching and related activities.
868 This is just like the C</pattern/> search, except that it matches only
869 once between calls to the reset() operator. This is a useful
870 optimization when you want to see only the first occurrence of
871 something in each file of a set of files, for instance. Only C<??>
872 patterns local to the current package are reset.
876 # blank line between header and body
879 reset if eof; # clear ?? status for next file
882 This usage is vaguely deprecated, which means it just might possibly
883 be removed in some distant future version of Perl, perhaps somewhere
884 around the year 2168.
886 =item m/PATTERN/cgimosx
888 =item /PATTERN/cgimosx
890 Searches a string for a pattern match, and in scalar context returns
891 true if it succeeds, false if it fails. If no string is specified
892 via the C<=~> or C<!~> operator, the $_ string is searched. (The
893 string specified with C<=~> need not be an lvalue--it may be the
894 result of an expression evaluation, but remember the C<=~> binds
895 rather tightly.) See also L<perlre>. See L<perllocale> for
896 discussion of additional considerations that apply when C<use locale>
901 c Do not reset search position on a failed match when /g is in effect.
902 g Match globally, i.e., find all occurrences.
903 i Do case-insensitive pattern matching.
904 m Treat string as multiple lines.
905 o Compile pattern only once.
906 s Treat string as single line.
907 x Use extended regular expressions.
909 If "/" is the delimiter then the initial C<m> is optional. With the C<m>
910 you can use any pair of non-alphanumeric, non-whitespace characters
911 as delimiters. This is particularly useful for matching path names
912 that contain "/", to avoid LTS (leaning toothpick syndrome). If "?" is
913 the delimiter, then the match-only-once rule of C<?PATTERN?> applies.
914 If "'" is the delimiter, no interpolation is performed on the PATTERN.
916 PATTERN may contain variables, which will be interpolated (and the
917 pattern recompiled) every time the pattern search is evaluated, except
918 for when the delimiter is a single quote. (Note that C<$(>, C<$)>, and
919 C<$|> are not interpolated because they look like end-of-string tests.)
920 If you want such a pattern to be compiled only once, add a C</o> after
921 the trailing delimiter. This avoids expensive run-time recompilations,
922 and is useful when the value you are interpolating won't change over
923 the life of the script. However, mentioning C</o> constitutes a promise
924 that you won't change the variables in the pattern. If you change them,
925 Perl won't even notice. See also L<"qr/STRING/imosx">.
927 If the PATTERN evaluates to the empty string, the last
928 I<successfully> matched regular expression is used instead. In this
929 case, only the C<g> and C<c> flags on the empty pattern is honoured -
930 the other flags are taken from the original pattern. If no match has
931 previously succeeded, this will (silently) act instead as a genuine
932 empty pattern (which will always match).
934 Note that it's possible to confuse Perl into thinking C<//> (the empty
935 regex) is really C<//> (the defined-or operator). Perl is usually pretty
936 good about this, but some pathological cases might trigger this, such as
937 C<$a///> (is that C<($a) / (//)> or C<$a // />?) and C<print $fh //>
938 (C<print $fh(//> or C<print($fh //>?). In all of these examples, Perl
939 will assume you meant defined-or. If you meant the empty regex, just
940 use parentheses or spaces to disambiguate, or even prefix the empty
941 regex with an C<m> (so C<//> becomes C<m//>).
943 If the C</g> option is not used, C<m//> in list context returns a
944 list consisting of the subexpressions matched by the parentheses in the
945 pattern, i.e., (C<$1>, C<$2>, C<$3>...). (Note that here C<$1> etc. are
946 also set, and that this differs from Perl 4's behavior.) When there are
947 no parentheses in the pattern, the return value is the list C<(1)> for
948 success. With or without parentheses, an empty list is returned upon
953 open(TTY, '/dev/tty');
954 <TTY> =~ /^y/i && foo(); # do foo if desired
956 if (/Version: *([0-9.]*)/) { $version = $1; }
958 next if m#^/usr/spool/uucp#;
963 print if /$arg/o; # compile only once
966 if (($F1, $F2, $Etc) = ($foo =~ /^(\S+)\s+(\S+)\s*(.*)/))
968 This last example splits $foo into the first two words and the
969 remainder of the line, and assigns those three fields to $F1, $F2, and
970 $Etc. The conditional is true if any variables were assigned, i.e., if
973 The C</g> modifier specifies global pattern matching--that is,
974 matching as many times as possible within the string. How it behaves
975 depends on the context. In list context, it returns a list of the
976 substrings matched by any capturing parentheses in the regular
977 expression. If there are no parentheses, it returns a list of all
978 the matched strings, as if there were parentheses around the whole
981 In scalar context, each execution of C<m//g> finds the next match,
982 returning true if it matches, and false if there is no further match.
983 The position after the last match can be read or set using the pos()
984 function; see L<perlfunc/pos>. A failed match normally resets the
985 search position to the beginning of the string, but you can avoid that
986 by adding the C</c> modifier (e.g. C<m//gc>). Modifying the target
987 string also resets the search position.
989 You can intermix C<m//g> matches with C<m/\G.../g>, where C<\G> is a
990 zero-width assertion that matches the exact position where the previous
991 C<m//g>, if any, left off. Without the C</g> modifier, the C<\G> assertion
992 still anchors at pos(), but the match is of course only attempted once.
993 Using C<\G> without C</g> on a target string that has not previously had a
994 C</g> match applied to it is the same as using the C<\A> assertion to match
995 the beginning of the string. Note also that, currently, C<\G> is only
996 properly supported when anchored at the very beginning of the pattern.
1001 ($one,$five,$fifteen) = (`uptime` =~ /(\d+\.\d+)/g);
1005 while (defined($paragraph = <>)) {
1006 while ($paragraph =~ /[a-z]['")]*[.!?]+['")]*\s/g) {
1010 print "$sentences\n";
1012 # using m//gc with \G
1016 print $1 while /(o)/gc; print "', pos=", pos, "\n";
1018 print $1 if /\G(q)/gc; print "', pos=", pos, "\n";
1020 print $1 while /(p)/gc; print "', pos=", pos, "\n";
1022 print "Final: '$1', pos=",pos,"\n" if /\G(.)/;
1024 The last example should print:
1034 Notice that the final match matched C<q> instead of C<p>, which a match
1035 without the C<\G> anchor would have done. Also note that the final match
1036 did not update C<pos> -- C<pos> is only updated on a C</g> match. If the
1037 final match did indeed match C<p>, it's a good bet that you're running an
1038 older (pre-5.6.0) Perl.
1040 A useful idiom for C<lex>-like scanners is C</\G.../gc>. You can
1041 combine several regexps like this to process a string part-by-part,
1042 doing different actions depending on which regexp matched. Each
1043 regexp tries to match where the previous one leaves off.
1046 $url = new URI::URL "http://www/"; die if $url eq "xXx";
1050 print(" digits"), redo LOOP if /\G\d+\b[,.;]?\s*/gc;
1051 print(" lowercase"), redo LOOP if /\G[a-z]+\b[,.;]?\s*/gc;
1052 print(" UPPERCASE"), redo LOOP if /\G[A-Z]+\b[,.;]?\s*/gc;
1053 print(" Capitalized"), redo LOOP if /\G[A-Z][a-z]+\b[,.;]?\s*/gc;
1054 print(" MiXeD"), redo LOOP if /\G[A-Za-z]+\b[,.;]?\s*/gc;
1055 print(" alphanumeric"), redo LOOP if /\G[A-Za-z0-9]+\b[,.;]?\s*/gc;
1056 print(" line-noise"), redo LOOP if /\G[^A-Za-z0-9]+/gc;
1057 print ". That's all!\n";
1060 Here is the output (split into several lines):
1062 line-noise lowercase line-noise lowercase UPPERCASE line-noise
1063 UPPERCASE line-noise lowercase line-noise lowercase line-noise
1064 lowercase lowercase line-noise lowercase lowercase line-noise
1065 MiXeD line-noise. That's all!
1071 A single-quoted, literal string. A backslash represents a backslash
1072 unless followed by the delimiter or another backslash, in which case
1073 the delimiter or backslash is interpolated.
1075 $foo = q!I said, "You said, 'She said it.'"!;
1076 $bar = q('This is it.');
1077 $baz = '\n'; # a two-character string
1083 A double-quoted, interpolated string.
1086 (*** The previous line contains the naughty word "$1".\n)
1087 if /\b(tcl|java|python)\b/i; # :-)
1088 $baz = "\n"; # a one-character string
1090 =item qr/STRING/imosx
1092 This operator quotes (and possibly compiles) its I<STRING> as a regular
1093 expression. I<STRING> is interpolated the same way as I<PATTERN>
1094 in C<m/PATTERN/>. If "'" is used as the delimiter, no interpolation
1095 is done. Returns a Perl value which may be used instead of the
1096 corresponding C</STRING/imosx> expression.
1100 $rex = qr/my.STRING/is;
1107 The result may be used as a subpattern in a match:
1110 $string =~ /foo${re}bar/; # can be interpolated in other patterns
1111 $string =~ $re; # or used standalone
1112 $string =~ /$re/; # or this way
1114 Since Perl may compile the pattern at the moment of execution of qr()
1115 operator, using qr() may have speed advantages in some situations,
1116 notably if the result of qr() is used standalone:
1119 my $patterns = shift;
1120 my @compiled = map qr/$_/i, @$patterns;
1123 foreach my $pat (@compiled) {
1124 $success = 1, last if /$pat/;
1130 Precompilation of the pattern into an internal representation at
1131 the moment of qr() avoids a need to recompile the pattern every
1132 time a match C</$pat/> is attempted. (Perl has many other internal
1133 optimizations, but none would be triggered in the above example if
1134 we did not use qr() operator.)
1138 i Do case-insensitive pattern matching.
1139 m Treat string as multiple lines.
1140 o Compile pattern only once.
1141 s Treat string as single line.
1142 x Use extended regular expressions.
1144 See L<perlre> for additional information on valid syntax for STRING, and
1145 for a detailed look at the semantics of regular expressions.
1151 A string which is (possibly) interpolated and then executed as a
1152 system command with C</bin/sh> or its equivalent. Shell wildcards,
1153 pipes, and redirections will be honored. The collected standard
1154 output of the command is returned; standard error is unaffected. In
1155 scalar context, it comes back as a single (potentially multi-line)
1156 string, or undef if the command failed. In list context, returns a
1157 list of lines (however you've defined lines with $/ or
1158 $INPUT_RECORD_SEPARATOR), or an empty list if the command failed.
1160 Because backticks do not affect standard error, use shell file descriptor
1161 syntax (assuming the shell supports this) if you care to address this.
1162 To capture a command's STDERR and STDOUT together:
1164 $output = `cmd 2>&1`;
1166 To capture a command's STDOUT but discard its STDERR:
1168 $output = `cmd 2>/dev/null`;
1170 To capture a command's STDERR but discard its STDOUT (ordering is
1173 $output = `cmd 2>&1 1>/dev/null`;
1175 To exchange a command's STDOUT and STDERR in order to capture the STDERR
1176 but leave its STDOUT to come out the old STDERR:
1178 $output = `cmd 3>&1 1>&2 2>&3 3>&-`;
1180 To read both a command's STDOUT and its STDERR separately, it's easiest
1181 and safest to redirect them separately to files, and then read from those
1182 files when the program is done:
1184 system("program args 1>/tmp/program.stdout 2>/tmp/program.stderr");
1186 Using single-quote as a delimiter protects the command from Perl's
1187 double-quote interpolation, passing it on to the shell instead:
1189 $perl_info = qx(ps $$); # that's Perl's $$
1190 $shell_info = qx'ps $$'; # that's the new shell's $$
1192 How that string gets evaluated is entirely subject to the command
1193 interpreter on your system. On most platforms, you will have to protect
1194 shell metacharacters if you want them treated literally. This is in
1195 practice difficult to do, as it's unclear how to escape which characters.
1196 See L<perlsec> for a clean and safe example of a manual fork() and exec()
1197 to emulate backticks safely.
1199 On some platforms (notably DOS-like ones), the shell may not be
1200 capable of dealing with multiline commands, so putting newlines in
1201 the string may not get you what you want. You may be able to evaluate
1202 multiple commands in a single line by separating them with the command
1203 separator character, if your shell supports that (e.g. C<;> on many Unix
1204 shells; C<&> on the Windows NT C<cmd> shell).
1206 Beginning with v5.6.0, Perl will attempt to flush all files opened for
1207 output before starting the child process, but this may not be supported
1208 on some platforms (see L<perlport>). To be safe, you may need to set
1209 C<$|> ($AUTOFLUSH in English) or call the C<autoflush()> method of
1210 C<IO::Handle> on any open handles.
1212 Beware that some command shells may place restrictions on the length
1213 of the command line. You must ensure your strings don't exceed this
1214 limit after any necessary interpolations. See the platform-specific
1215 release notes for more details about your particular environment.
1217 Using this operator can lead to programs that are difficult to port,
1218 because the shell commands called vary between systems, and may in
1219 fact not be present at all. As one example, the C<type> command under
1220 the POSIX shell is very different from the C<type> command under DOS.
1221 That doesn't mean you should go out of your way to avoid backticks
1222 when they're the right way to get something done. Perl was made to be
1223 a glue language, and one of the things it glues together is commands.
1224 Just understand what you're getting yourself into.
1226 See L<"I/O Operators"> for more discussion.
1230 Evaluates to a list of the words extracted out of STRING, using embedded
1231 whitespace as the word delimiters. It can be understood as being roughly
1234 split(' ', q/STRING/);
1236 the differences being that it generates a real list at compile time, and
1237 in scalar context it returns the last element in the list. So
1242 is semantically equivalent to the list:
1246 Some frequently seen examples:
1248 use POSIX qw( setlocale localeconv )
1249 @EXPORT = qw( foo bar baz );
1251 A common mistake is to try to separate the words with comma or to
1252 put comments into a multi-line C<qw>-string. For this reason, the
1253 C<use warnings> pragma and the B<-w> switch (that is, the C<$^W> variable)
1254 produces warnings if the STRING contains the "," or the "#" character.
1256 =item s/PATTERN/REPLACEMENT/egimosx
1258 Searches a string for a pattern, and if found, replaces that pattern
1259 with the replacement text and returns the number of substitutions
1260 made. Otherwise it returns false (specifically, the empty string).
1262 If no string is specified via the C<=~> or C<!~> operator, the C<$_>
1263 variable is searched and modified. (The string specified with C<=~> must
1264 be scalar variable, an array element, a hash element, or an assignment
1265 to one of those, i.e., an lvalue.)
1267 If the delimiter chosen is a single quote, no interpolation is
1268 done on either the PATTERN or the REPLACEMENT. Otherwise, if the
1269 PATTERN contains a $ that looks like a variable rather than an
1270 end-of-string test, the variable will be interpolated into the pattern
1271 at run-time. If you want the pattern compiled only once the first time
1272 the variable is interpolated, use the C</o> option. If the pattern
1273 evaluates to the empty string, the last successfully executed regular
1274 expression is used instead. See L<perlre> for further explanation on these.
1275 See L<perllocale> for discussion of additional considerations that apply
1276 when C<use locale> is in effect.
1280 e Evaluate the right side as an expression.
1281 g Replace globally, i.e., all occurrences.
1282 i Do case-insensitive pattern matching.
1283 m Treat string as multiple lines.
1284 o Compile pattern only once.
1285 s Treat string as single line.
1286 x Use extended regular expressions.
1288 Any non-alphanumeric, non-whitespace delimiter may replace the
1289 slashes. If single quotes are used, no interpretation is done on the
1290 replacement string (the C</e> modifier overrides this, however). Unlike
1291 Perl 4, Perl 5 treats backticks as normal delimiters; the replacement
1292 text is not evaluated as a command. If the
1293 PATTERN is delimited by bracketing quotes, the REPLACEMENT has its own
1294 pair of quotes, which may or may not be bracketing quotes, e.g.,
1295 C<s(foo)(bar)> or C<< s<foo>/bar/ >>. A C</e> will cause the
1296 replacement portion to be treated as a full-fledged Perl expression
1297 and evaluated right then and there. It is, however, syntax checked at
1298 compile-time. A second C<e> modifier will cause the replacement portion
1299 to be C<eval>ed before being run as a Perl expression.
1303 s/\bgreen\b/mauve/g; # don't change wintergreen
1305 $path =~ s|/usr/bin|/usr/local/bin|;
1307 s/Login: $foo/Login: $bar/; # run-time pattern
1309 ($foo = $bar) =~ s/this/that/; # copy first, then change
1311 $count = ($paragraph =~ s/Mister\b/Mr./g); # get change-count
1314 s/\d+/$&*2/e; # yields 'abc246xyz'
1315 s/\d+/sprintf("%5d",$&)/e; # yields 'abc 246xyz'
1316 s/\w/$& x 2/eg; # yields 'aabbcc 224466xxyyzz'
1318 s/%(.)/$percent{$1}/g; # change percent escapes; no /e
1319 s/%(.)/$percent{$1} || $&/ge; # expr now, so /e
1320 s/^=(\w+)/&pod($1)/ge; # use function call
1322 # expand variables in $_, but dynamics only, using
1323 # symbolic dereferencing
1326 # Add one to the value of any numbers in the string
1329 # This will expand any embedded scalar variable
1330 # (including lexicals) in $_ : First $1 is interpolated
1331 # to the variable name, and then evaluated
1334 # Delete (most) C comments.
1336 /\* # Match the opening delimiter.
1337 .*? # Match a minimal number of characters.
1338 \*/ # Match the closing delimiter.
1341 s/^\s*(.*?)\s*$/$1/; # trim white space in $_, expensively
1343 for ($variable) { # trim white space in $variable, cheap
1348 s/([^ ]*) *([^ ]*)/$2 $1/; # reverse 1st two fields
1350 Note the use of $ instead of \ in the last example. Unlike
1351 B<sed>, we use the \<I<digit>> form in only the left hand side.
1352 Anywhere else it's $<I<digit>>.
1354 Occasionally, you can't use just a C</g> to get all the changes
1355 to occur that you might want. Here are two common cases:
1357 # put commas in the right places in an integer
1358 1 while s/(\d)(\d\d\d)(?!\d)/$1,$2/g;
1360 # expand tabs to 8-column spacing
1361 1 while s/\t+/' ' x (length($&)*8 - length($`)%8)/e;
1363 =item tr/SEARCHLIST/REPLACEMENTLIST/cds
1365 =item y/SEARCHLIST/REPLACEMENTLIST/cds
1367 Transliterates all occurrences of the characters found in the search list
1368 with the corresponding character in the replacement list. It returns
1369 the number of characters replaced or deleted. If no string is
1370 specified via the =~ or !~ operator, the $_ string is transliterated. (The
1371 string specified with =~ must be a scalar variable, an array element, a
1372 hash element, or an assignment to one of those, i.e., an lvalue.)
1374 A character range may be specified with a hyphen, so C<tr/A-J/0-9/>
1375 does the same replacement as C<tr/ACEGIBDFHJ/0246813579/>.
1376 For B<sed> devotees, C<y> is provided as a synonym for C<tr>. If the
1377 SEARCHLIST is delimited by bracketing quotes, the REPLACEMENTLIST has
1378 its own pair of quotes, which may or may not be bracketing quotes,
1379 e.g., C<tr[A-Z][a-z]> or C<tr(+\-*/)/ABCD/>.
1381 Note that C<tr> does B<not> do regular expression character classes
1382 such as C<\d> or C<[:lower:]>. The <tr> operator is not equivalent to
1383 the tr(1) utility. If you want to map strings between lower/upper
1384 cases, see L<perlfunc/lc> and L<perlfunc/uc>, and in general consider
1385 using the C<s> operator if you need regular expressions.
1387 Note also that the whole range idea is rather unportable between
1388 character sets--and even within character sets they may cause results
1389 you probably didn't expect. A sound principle is to use only ranges
1390 that begin from and end at either alphabets of equal case (a-e, A-E),
1391 or digits (0-4). Anything else is unsafe. If in doubt, spell out the
1392 character sets in full.
1396 c Complement the SEARCHLIST.
1397 d Delete found but unreplaced characters.
1398 s Squash duplicate replaced characters.
1400 If the C</c> modifier is specified, the SEARCHLIST character set
1401 is complemented. If the C</d> modifier is specified, any characters
1402 specified by SEARCHLIST not found in REPLACEMENTLIST are deleted.
1403 (Note that this is slightly more flexible than the behavior of some
1404 B<tr> programs, which delete anything they find in the SEARCHLIST,
1405 period.) If the C</s> modifier is specified, sequences of characters
1406 that were transliterated to the same character are squashed down
1407 to a single instance of the character.
1409 If the C</d> modifier is used, the REPLACEMENTLIST is always interpreted
1410 exactly as specified. Otherwise, if the REPLACEMENTLIST is shorter
1411 than the SEARCHLIST, the final character is replicated till it is long
1412 enough. If the REPLACEMENTLIST is empty, the SEARCHLIST is replicated.
1413 This latter is useful for counting characters in a class or for
1414 squashing character sequences in a class.
1418 $ARGV[1] =~ tr/A-Z/a-z/; # canonicalize to lower case
1420 $cnt = tr/*/*/; # count the stars in $_
1422 $cnt = $sky =~ tr/*/*/; # count the stars in $sky
1424 $cnt = tr/0-9//; # count the digits in $_
1426 tr/a-zA-Z//s; # bookkeeper -> bokeper
1428 ($HOST = $host) =~ tr/a-z/A-Z/;
1430 tr/a-zA-Z/ /cs; # change non-alphas to single space
1433 [\000-\177]; # delete 8th bit
1435 If multiple transliterations are given for a character, only the
1440 will transliterate any A to X.
1442 Because the transliteration table is built at compile time, neither
1443 the SEARCHLIST nor the REPLACEMENTLIST are subjected to double quote
1444 interpolation. That means that if you want to use variables, you
1447 eval "tr/$oldlist/$newlist/";
1450 eval "tr/$oldlist/$newlist/, 1" or die $@;
1454 A line-oriented form of quoting is based on the shell "here-document"
1455 syntax. Following a C<< << >> you specify a string to terminate
1456 the quoted material, and all lines following the current line down to
1457 the terminating string are the value of the item. The terminating
1458 string may be either an identifier (a word), or some quoted text. If
1459 quoted, the type of quotes you use determines the treatment of the
1460 text, just as in regular quoting. An unquoted identifier works like
1461 double quotes. There must be no space between the C<< << >> and
1462 the identifier, unless the identifier is quoted. (If you put a space it
1463 will be treated as a null identifier, which is valid, and matches the first
1464 empty line.) The terminating string must appear by itself (unquoted and
1465 with no surrounding whitespace) on the terminating line.
1468 The price is $Price.
1471 print << "EOF"; # same as above
1472 The price is $Price.
1475 print << `EOC`; # execute commands
1480 print <<"foo", <<"bar"; # you can stack them
1486 myfunc(<< "THIS", 23, <<'THAT');
1493 Just don't forget that you have to put a semicolon on the end
1494 to finish the statement, as Perl doesn't know you're not going to
1502 If you want your here-docs to be indented with the
1503 rest of the code, you'll need to remove leading whitespace
1504 from each line manually:
1506 ($quote = <<'FINIS') =~ s/^\s+//gm;
1507 The Road goes ever on and on,
1508 down from the door where it began.
1511 If you use a here-doc within a delimited construct, such as in C<s///eg>,
1512 the quoted material must come on the lines following the final delimiter.
1527 If the terminating identifier is on the last line of the program, you
1528 must be sure there is a newline after it; otherwise, Perl will give the
1529 warning B<Can't find string terminator "END" anywhere before EOF...>.
1531 Additionally, the quoting rules for the identifier are not related to
1532 Perl's quoting rules -- C<q()>, C<qq()>, and the like are not supported
1533 in place of C<''> and C<"">, and the only interpolation is for backslashing
1534 the quoting character:
1536 print << "abc\"def";
1540 Finally, quoted strings cannot span multiple lines. The general rule is
1541 that the identifier must be a string literal. Stick with that, and you
1546 =head2 Gory details of parsing quoted constructs
1548 When presented with something that might have several different
1549 interpretations, Perl uses the B<DWIM> (that's "Do What I Mean")
1550 principle to pick the most probable interpretation. This strategy
1551 is so successful that Perl programmers often do not suspect the
1552 ambivalence of what they write. But from time to time, Perl's
1553 notions differ substantially from what the author honestly meant.
1555 This section hopes to clarify how Perl handles quoted constructs.
1556 Although the most common reason to learn this is to unravel labyrinthine
1557 regular expressions, because the initial steps of parsing are the
1558 same for all quoting operators, they are all discussed together.
1560 The most important Perl parsing rule is the first one discussed
1561 below: when processing a quoted construct, Perl first finds the end
1562 of that construct, then interprets its contents. If you understand
1563 this rule, you may skip the rest of this section on the first
1564 reading. The other rules are likely to contradict the user's
1565 expectations much less frequently than this first one.
1567 Some passes discussed below are performed concurrently, but because
1568 their results are the same, we consider them individually. For different
1569 quoting constructs, Perl performs different numbers of passes, from
1570 one to five, but these passes are always performed in the same order.
1574 =item Finding the end
1576 The first pass is finding the end of the quoted construct, whether
1577 it be a multicharacter delimiter C<"\nEOF\n"> in the C<<<EOF>
1578 construct, a C</> that terminates a C<qq//> construct, a C<]> which
1579 terminates C<qq[]> construct, or a C<< > >> which terminates a
1580 fileglob started with C<< < >>.
1582 When searching for single-character non-pairing delimiters, such
1583 as C</>, combinations of C<\\> and C<\/> are skipped. However,
1584 when searching for single-character pairing delimiter like C<[>,
1585 combinations of C<\\>, C<\]>, and C<\[> are all skipped, and nested
1586 C<[>, C<]> are skipped as well. When searching for multicharacter
1587 delimiters, nothing is skipped.
1589 For constructs with three-part delimiters (C<s///>, C<y///>, and
1590 C<tr///>), the search is repeated once more.
1592 During this search no attention is paid to the semantics of the construct.
1595 "$hash{"$foo/$bar"}"
1600 bar # NOT a comment, this slash / terminated m//!
1603 do not form legal quoted expressions. The quoted part ends on the
1604 first C<"> and C</>, and the rest happens to be a syntax error.
1605 Because the slash that terminated C<m//> was followed by a C<SPACE>,
1606 the example above is not C<m//x>, but rather C<m//> with no C</x>
1607 modifier. So the embedded C<#> is interpreted as a literal C<#>.
1609 =item Removal of backslashes before delimiters
1611 During the second pass, text between the starting and ending
1612 delimiters is copied to a safe location, and the C<\> is removed
1613 from combinations consisting of C<\> and delimiter--or delimiters,
1614 meaning both starting and ending delimiters will should these differ.
1615 This removal does not happen for multi-character delimiters.
1616 Note that the combination C<\\> is left intact, just as it was.
1618 Starting from this step no information about the delimiters is
1623 The next step is interpolation in the text obtained, which is now
1624 delimiter-independent. There are four different cases.
1628 =item C<<<'EOF'>, C<m''>, C<s'''>, C<tr///>, C<y///>
1630 No interpolation is performed.
1634 The only interpolation is removal of C<\> from pairs C<\\>.
1636 =item C<"">, C<``>, C<qq//>, C<qx//>, C<< <file*glob> >>
1638 C<\Q>, C<\U>, C<\u>, C<\L>, C<\l> (possibly paired with C<\E>) are
1639 converted to corresponding Perl constructs. Thus, C<"$foo\Qbaz$bar">
1640 is converted to C<$foo . (quotemeta("baz" . $bar))> internally.
1641 The other combinations are replaced with appropriate expansions.
1643 Let it be stressed that I<whatever falls between C<\Q> and C<\E>>
1644 is interpolated in the usual way. Something like C<"\Q\\E"> has
1645 no C<\E> inside. instead, it has C<\Q>, C<\\>, and C<E>, so the
1646 result is the same as for C<"\\\\E">. As a general rule, backslashes
1647 between C<\Q> and C<\E> may lead to counterintuitive results. So,
1648 C<"\Q\t\E"> is converted to C<quotemeta("\t")>, which is the same
1649 as C<"\\\t"> (since TAB is not alphanumeric). Note also that:
1654 may be closer to the conjectural I<intention> of the writer of C<"\Q\t\E">.
1656 Interpolated scalars and arrays are converted internally to the C<join> and
1657 C<.> catenation operations. Thus, C<"$foo XXX '@arr'"> becomes:
1659 $foo . " XXX '" . (join $", @arr) . "'";
1661 All operations above are performed simultaneously, left to right.
1663 Because the result of C<"\Q STRING \E"> has all metacharacters
1664 quoted, there is no way to insert a literal C<$> or C<@> inside a
1665 C<\Q\E> pair. If protected by C<\>, C<$> will be quoted to became
1666 C<"\\\$">; if not, it is interpreted as the start of an interpolated
1669 Note also that the interpolation code needs to make a decision on
1670 where the interpolated scalar ends. For instance, whether
1671 C<< "a $b -> {c}" >> really means:
1673 "a " . $b . " -> {c}";
1679 Most of the time, the longest possible text that does not include
1680 spaces between components and which contains matching braces or
1681 brackets. because the outcome may be determined by voting based
1682 on heuristic estimators, the result is not strictly predictable.
1683 Fortunately, it's usually correct for ambiguous cases.
1685 =item C<?RE?>, C</RE/>, C<m/RE/>, C<s/RE/foo/>,
1687 Processing of C<\Q>, C<\U>, C<\u>, C<\L>, C<\l>, and interpolation
1688 happens (almost) as with C<qq//> constructs, but the substitution
1689 of C<\> followed by RE-special chars (including C<\>) is not
1690 performed. Moreover, inside C<(?{BLOCK})>, C<(?# comment )>, and
1691 a C<#>-comment in a C<//x>-regular expression, no processing is
1692 performed whatsoever. This is the first step at which the presence
1693 of the C<//x> modifier is relevant.
1695 Interpolation has several quirks: C<$|>, C<$(>, and C<$)> are not
1696 interpolated, and constructs C<$var[SOMETHING]> are voted (by several
1697 different estimators) to be either an array element or C<$var>
1698 followed by an RE alternative. This is where the notation
1699 C<${arr[$bar]}> comes handy: C</${arr[0-9]}/> is interpreted as
1700 array element C<-9>, not as a regular expression from the variable
1701 C<$arr> followed by a digit, which would be the interpretation of
1702 C</$arr[0-9]/>. Since voting among different estimators may occur,
1703 the result is not predictable.
1705 It is at this step that C<\1> is begrudgingly converted to C<$1> in
1706 the replacement text of C<s///> to correct the incorrigible
1707 I<sed> hackers who haven't picked up the saner idiom yet. A warning
1708 is emitted if the C<use warnings> pragma or the B<-w> command-line flag
1709 (that is, the C<$^W> variable) was set.
1711 The lack of processing of C<\\> creates specific restrictions on
1712 the post-processed text. If the delimiter is C</>, one cannot get
1713 the combination C<\/> into the result of this step. C</> will
1714 finish the regular expression, C<\/> will be stripped to C</> on
1715 the previous step, and C<\\/> will be left as is. Because C</> is
1716 equivalent to C<\/> inside a regular expression, this does not
1717 matter unless the delimiter happens to be character special to the
1718 RE engine, such as in C<s*foo*bar*>, C<m[foo]>, or C<?foo?>; or an
1719 alphanumeric char, as in:
1723 In the RE above, which is intentionally obfuscated for illustration, the
1724 delimiter is C<m>, the modifier is C<mx>, and after backslash-removal the
1725 RE is the same as for C<m/ ^ a \s* b /mx>. There's more than one
1726 reason you're encouraged to restrict your delimiters to non-alphanumeric,
1727 non-whitespace choices.
1731 This step is the last one for all constructs except regular expressions,
1732 which are processed further.
1734 =item Interpolation of regular expressions
1736 Previous steps were performed during the compilation of Perl code,
1737 but this one happens at run time--although it may be optimized to
1738 be calculated at compile time if appropriate. After preprocessing
1739 described above, and possibly after evaluation if catenation,
1740 joining, casing translation, or metaquoting are involved, the
1741 resulting I<string> is passed to the RE engine for compilation.
1743 Whatever happens in the RE engine might be better discussed in L<perlre>,
1744 but for the sake of continuity, we shall do so here.
1746 This is another step where the presence of the C<//x> modifier is
1747 relevant. The RE engine scans the string from left to right and
1748 converts it to a finite automaton.
1750 Backslashed characters are either replaced with corresponding
1751 literal strings (as with C<\{>), or else they generate special nodes
1752 in the finite automaton (as with C<\b>). Characters special to the
1753 RE engine (such as C<|>) generate corresponding nodes or groups of
1754 nodes. C<(?#...)> comments are ignored. All the rest is either
1755 converted to literal strings to match, or else is ignored (as is
1756 whitespace and C<#>-style comments if C<//x> is present).
1758 Parsing of the bracketed character class construct, C<[...]>, is
1759 rather different than the rule used for the rest of the pattern.
1760 The terminator of this construct is found using the same rules as
1761 for finding the terminator of a C<{}>-delimited construct, the only
1762 exception being that C<]> immediately following C<[> is treated as
1763 though preceded by a backslash. Similarly, the terminator of
1764 C<(?{...})> is found using the same rules as for finding the
1765 terminator of a C<{}>-delimited construct.
1767 It is possible to inspect both the string given to RE engine and the
1768 resulting finite automaton. See the arguments C<debug>/C<debugcolor>
1769 in the C<use L<re>> pragma, as well as Perl's B<-Dr> command-line
1770 switch documented in L<perlrun/"Command Switches">.
1772 =item Optimization of regular expressions
1774 This step is listed for completeness only. Since it does not change
1775 semantics, details of this step are not documented and are subject
1776 to change without notice. This step is performed over the finite
1777 automaton that was generated during the previous pass.
1779 It is at this stage that C<split()> silently optimizes C</^/> to
1784 =head2 I/O Operators
1786 There are several I/O operators you should know about.
1788 A string enclosed by backticks (grave accents) first undergoes
1789 double-quote interpolation. It is then interpreted as an external
1790 command, and the output of that command is the value of the
1791 backtick string, like in a shell. In scalar context, a single string
1792 consisting of all output is returned. In list context, a list of
1793 values is returned, one per line of output. (You can set C<$/> to use
1794 a different line terminator.) The command is executed each time the
1795 pseudo-literal is evaluated. The status value of the command is
1796 returned in C<$?> (see L<perlvar> for the interpretation of C<$?>).
1797 Unlike in B<csh>, no translation is done on the return data--newlines
1798 remain newlines. Unlike in any of the shells, single quotes do not
1799 hide variable names in the command from interpretation. To pass a
1800 literal dollar-sign through to the shell you need to hide it with a
1801 backslash. The generalized form of backticks is C<qx//>. (Because
1802 backticks always undergo shell expansion as well, see L<perlsec> for
1805 In scalar context, evaluating a filehandle in angle brackets yields
1806 the next line from that file (the newline, if any, included), or
1807 C<undef> at end-of-file or on error. When C<$/> is set to C<undef>
1808 (sometimes known as file-slurp mode) and the file is empty, it
1809 returns C<''> the first time, followed by C<undef> subsequently.
1811 Ordinarily you must assign the returned value to a variable, but
1812 there is one situation where an automatic assignment happens. If
1813 and only if the input symbol is the only thing inside the conditional
1814 of a C<while> statement (even if disguised as a C<for(;;)> loop),
1815 the value is automatically assigned to the global variable $_,
1816 destroying whatever was there previously. (This may seem like an
1817 odd thing to you, but you'll use the construct in almost every Perl
1818 script you write.) The $_ variable is not implicitly localized.
1819 You'll have to put a C<local $_;> before the loop if you want that
1822 The following lines are equivalent:
1824 while (defined($_ = <STDIN>)) { print; }
1825 while ($_ = <STDIN>) { print; }
1826 while (<STDIN>) { print; }
1827 for (;<STDIN>;) { print; }
1828 print while defined($_ = <STDIN>);
1829 print while ($_ = <STDIN>);
1830 print while <STDIN>;
1832 This also behaves similarly, but avoids $_ :
1834 while (my $line = <STDIN>) { print $line }
1836 In these loop constructs, the assigned value (whether assignment
1837 is automatic or explicit) is then tested to see whether it is
1838 defined. The defined test avoids problems where line has a string
1839 value that would be treated as false by Perl, for example a "" or
1840 a "0" with no trailing newline. If you really mean for such values
1841 to terminate the loop, they should be tested for explicitly:
1843 while (($_ = <STDIN>) ne '0') { ... }
1844 while (<STDIN>) { last unless $_; ... }
1846 In other boolean contexts, C<< <I<filehandle>> >> without an
1847 explicit C<defined> test or comparison elicit a warning if the
1848 C<use warnings> pragma or the B<-w>
1849 command-line switch (the C<$^W> variable) is in effect.
1851 The filehandles STDIN, STDOUT, and STDERR are predefined. (The
1852 filehandles C<stdin>, C<stdout>, and C<stderr> will also work except
1853 in packages, where they would be interpreted as local identifiers
1854 rather than global.) Additional filehandles may be created with
1855 the open() function, amongst others. See L<perlopentut> and
1856 L<perlfunc/open> for details on this.
1858 If a <FILEHANDLE> is used in a context that is looking for
1859 a list, a list comprising all input lines is returned, one line per
1860 list element. It's easy to grow to a rather large data space this
1861 way, so use with care.
1863 <FILEHANDLE> may also be spelled C<readline(*FILEHANDLE)>.
1864 See L<perlfunc/readline>.
1866 The null filehandle <> is special: it can be used to emulate the
1867 behavior of B<sed> and B<awk>. Input from <> comes either from
1868 standard input, or from each file listed on the command line. Here's
1869 how it works: the first time <> is evaluated, the @ARGV array is
1870 checked, and if it is empty, C<$ARGV[0]> is set to "-", which when opened
1871 gives you standard input. The @ARGV array is then processed as a list
1872 of filenames. The loop
1875 ... # code for each line
1878 is equivalent to the following Perl-like pseudo code:
1880 unshift(@ARGV, '-') unless @ARGV;
1881 while ($ARGV = shift) {
1884 ... # code for each line
1888 except that it isn't so cumbersome to say, and will actually work.
1889 It really does shift the @ARGV array and put the current filename
1890 into the $ARGV variable. It also uses filehandle I<ARGV>
1891 internally--<> is just a synonym for <ARGV>, which
1892 is magical. (The pseudo code above doesn't work because it treats
1893 <ARGV> as non-magical.)
1895 You can modify @ARGV before the first <> as long as the array ends up
1896 containing the list of filenames you really want. Line numbers (C<$.>)
1897 continue as though the input were one big happy file. See the example
1898 in L<perlfunc/eof> for how to reset line numbers on each file.
1900 If you want to set @ARGV to your own list of files, go right ahead.
1901 This sets @ARGV to all plain text files if no @ARGV was given:
1903 @ARGV = grep { -f && -T } glob('*') unless @ARGV;
1905 You can even set them to pipe commands. For example, this automatically
1906 filters compressed arguments through B<gzip>:
1908 @ARGV = map { /\.(gz|Z)$/ ? "gzip -dc < $_ |" : $_ } @ARGV;
1910 If you want to pass switches into your script, you can use one of the
1911 Getopts modules or put a loop on the front like this:
1913 while ($_ = $ARGV[0], /^-/) {
1916 if (/^-D(.*)/) { $debug = $1 }
1917 if (/^-v/) { $verbose++ }
1918 # ... # other switches
1922 # ... # code for each line
1925 The <> symbol will return C<undef> for end-of-file only once.
1926 If you call it again after this, it will assume you are processing another
1927 @ARGV list, and if you haven't set @ARGV, will read input from STDIN.
1929 If what the angle brackets contain is a simple scalar variable (e.g.,
1930 <$foo>), then that variable contains the name of the
1931 filehandle to input from, or its typeglob, or a reference to the
1937 If what's within the angle brackets is neither a filehandle nor a simple
1938 scalar variable containing a filehandle name, typeglob, or typeglob
1939 reference, it is interpreted as a filename pattern to be globbed, and
1940 either a list of filenames or the next filename in the list is returned,
1941 depending on context. This distinction is determined on syntactic
1942 grounds alone. That means C<< <$x> >> is always a readline() from
1943 an indirect handle, but C<< <$hash{key}> >> is always a glob().
1944 That's because $x is a simple scalar variable, but C<$hash{key}> is
1945 not--it's a hash element.
1947 One level of double-quote interpretation is done first, but you can't
1948 say C<< <$foo> >> because that's an indirect filehandle as explained
1949 in the previous paragraph. (In older versions of Perl, programmers
1950 would insert curly brackets to force interpretation as a filename glob:
1951 C<< <${foo}> >>. These days, it's considered cleaner to call the
1952 internal function directly as C<glob($foo)>, which is probably the right
1953 way to have done it in the first place.) For example:
1959 is roughly equivalent to:
1961 open(FOO, "echo *.c | tr -s ' \t\r\f' '\\012\\012\\012\\012'|");
1967 except that the globbing is actually done internally using the standard
1968 C<File::Glob> extension. Of course, the shortest way to do the above is:
1972 A (file)glob evaluates its (embedded) argument only when it is
1973 starting a new list. All values must be read before it will start
1974 over. In list context, this isn't important because you automatically
1975 get them all anyway. However, in scalar context the operator returns
1976 the next value each time it's called, or C<undef> when the list has
1977 run out. As with filehandle reads, an automatic C<defined> is
1978 generated when the glob occurs in the test part of a C<while>,
1979 because legal glob returns (e.g. a file called F<0>) would otherwise
1980 terminate the loop. Again, C<undef> is returned only once. So if
1981 you're expecting a single value from a glob, it is much better to
1984 ($file) = <blurch*>;
1990 because the latter will alternate between returning a filename and
1993 If you're trying to do variable interpolation, it's definitely better
1994 to use the glob() function, because the older notation can cause people
1995 to become confused with the indirect filehandle notation.
1997 @files = glob("$dir/*.[ch]");
1998 @files = glob($files[$i]);
2000 =head2 Constant Folding
2002 Like C, Perl does a certain amount of expression evaluation at
2003 compile time whenever it determines that all arguments to an
2004 operator are static and have no side effects. In particular, string
2005 concatenation happens at compile time between literals that don't do
2006 variable substitution. Backslash interpolation also happens at
2007 compile time. You can say
2009 'Now is the time for all' . "\n" .
2010 'good men to come to.'
2012 and this all reduces to one string internally. Likewise, if
2015 foreach $file (@filenames) {
2016 if (-s $file > 5 + 100 * 2**16) { }
2019 the compiler will precompute the number which that expression
2020 represents so that the interpreter won't have to.
2022 =head2 Bitwise String Operators
2024 Bitstrings of any size may be manipulated by the bitwise operators
2027 If the operands to a binary bitwise op are strings of different
2028 sizes, B<|> and B<^> ops act as though the shorter operand had
2029 additional zero bits on the right, while the B<&> op acts as though
2030 the longer operand were truncated to the length of the shorter.
2031 The granularity for such extension or truncation is one or more
2034 # ASCII-based examples
2035 print "j p \n" ^ " a h"; # prints "JAPH\n"
2036 print "JA" | " ph\n"; # prints "japh\n"
2037 print "japh\nJunk" & '_____'; # prints "JAPH\n";
2038 print 'p N$' ^ " E<H\n"; # prints "Perl\n";
2040 If you are intending to manipulate bitstrings, be certain that
2041 you're supplying bitstrings: If an operand is a number, that will imply
2042 a B<numeric> bitwise operation. You may explicitly show which type of
2043 operation you intend by using C<""> or C<0+>, as in the examples below.
2045 $foo = 150 | 105 ; # yields 255 (0x96 | 0x69 is 0xFF)
2046 $foo = '150' | 105 ; # yields 255
2047 $foo = 150 | '105'; # yields 255
2048 $foo = '150' | '105'; # yields string '155' (under ASCII)
2050 $baz = 0+$foo & 0+$bar; # both ops explicitly numeric
2051 $biz = "$foo" ^ "$bar"; # both ops explicitly stringy
2053 See L<perlfunc/vec> for information on how to manipulate individual bits
2056 =head2 Integer Arithmetic
2058 By default, Perl assumes that it must do most of its arithmetic in
2059 floating point. But by saying
2063 you may tell the compiler that it's okay to use integer operations
2064 (if it feels like it) from here to the end of the enclosing BLOCK.
2065 An inner BLOCK may countermand this by saying
2069 which lasts until the end of that BLOCK. Note that this doesn't
2070 mean everything is only an integer, merely that Perl may use integer
2071 operations if it is so inclined. For example, even under C<use
2072 integer>, if you take the C<sqrt(2)>, you'll still get C<1.4142135623731>
2075 Used on numbers, the bitwise operators ("&", "|", "^", "~", "<<",
2076 and ">>") always produce integral results. (But see also
2077 L<Bitwise String Operators>.) However, C<use integer> still has meaning for
2078 them. By default, their results are interpreted as unsigned integers, but
2079 if C<use integer> is in effect, their results are interpreted
2080 as signed integers. For example, C<~0> usually evaluates to a large
2081 integral value. However, C<use integer; ~0> is C<-1> on twos-complement
2084 =head2 Floating-point Arithmetic
2086 While C<use integer> provides integer-only arithmetic, there is no
2087 analogous mechanism to provide automatic rounding or truncation to a
2088 certain number of decimal places. For rounding to a certain number
2089 of digits, sprintf() or printf() is usually the easiest route.
2092 Floating-point numbers are only approximations to what a mathematician
2093 would call real numbers. There are infinitely more reals than floats,
2094 so some corners must be cut. For example:
2096 printf "%.20g\n", 123456789123456789;
2097 # produces 123456789123456784
2099 Testing for exact equality of floating-point equality or inequality is
2100 not a good idea. Here's a (relatively expensive) work-around to compare
2101 whether two floating-point numbers are equal to a particular number of
2102 decimal places. See Knuth, volume II, for a more robust treatment of
2106 my ($X, $Y, $POINTS) = @_;
2108 $tX = sprintf("%.${POINTS}g", $X);
2109 $tY = sprintf("%.${POINTS}g", $Y);
2113 The POSIX module (part of the standard perl distribution) implements
2114 ceil(), floor(), and other mathematical and trigonometric functions.
2115 The Math::Complex module (part of the standard perl distribution)
2116 defines mathematical functions that work on both the reals and the
2117 imaginary numbers. Math::Complex not as efficient as POSIX, but
2118 POSIX can't work with complex numbers.
2120 Rounding in financial applications can have serious implications, and
2121 the rounding method used should be specified precisely. In these
2122 cases, it probably pays not to trust whichever system rounding is
2123 being used by Perl, but to instead implement the rounding function you
2126 =head2 Bigger Numbers
2128 The standard Math::BigInt and Math::BigFloat modules provide
2129 variable-precision arithmetic and overloaded operators, although
2130 they're currently pretty slow. At the cost of some space and
2131 considerable speed, they avoid the normal pitfalls associated with
2132 limited-precision representations.
2135 $x = Math::BigInt->new('123456789123456789');
2138 # prints +15241578780673678515622620750190521
2140 There are several modules that let you calculate with (bound only by
2141 memory and cpu-time) unlimited or fixed precision. There are also
2142 some non-standard modules that provide faster implementations via
2143 external C libraries.
2145 Here is a short, but incomplete summary:
2147 Math::Fraction big, unlimited fractions like 9973 / 12967
2148 Math::String treat string sequences like numbers
2149 Math::FixedPrecision calculate with a fixed precision
2150 Math::Currency for currency calculations
2151 Bit::Vector manipulate bit vectors fast (uses C)
2152 Math::BigIntFast Bit::Vector wrapper for big numbers
2153 Math::Pari provides access to the Pari C library
2154 Math::BigInteger uses an external C library
2155 Math::Cephes uses external Cephes C library (no big numbers)
2156 Math::Cephes::Fraction fractions via the Cephes library
2157 Math::GMP another one using an external C library