3 perlfaq4 - Data Manipulation ($Revision: 8539 $)
7 This section of the FAQ answers questions related to manipulating
8 numbers, dates, strings, arrays, hashes, and miscellaneous data issues.
12 =head2 Why am I getting long decimals (eg, 19.9499999999999) instead of the numbers I should be getting (eg, 19.95)?
14 Internally, your computer represents floating-point numbers in binary.
15 Digital (as in powers of two) computers cannot store all numbers
16 exactly. Some real numbers lose precision in the process. This is a
17 problem with how computers store numbers and affects all computer
18 languages, not just Perl.
20 L<perlnumber> shows the gory details of number representations and
23 To limit the number of decimal places in your numbers, you can use the
24 printf or sprintf function. See the L<"Floating Point
25 Arithmetic"|perlop> for more details.
29 my $number = sprintf "%.2f", 10/3;
31 =head2 Why is int() broken?
33 Your C<int()> is most probably working just fine. It's the numbers that
34 aren't quite what you think.
36 First, see the answer to "Why am I getting long decimals
37 (eg, 19.9499999999999) instead of the numbers I should be getting
42 print int(0.6/0.2-2), "\n";
44 will in most computers print 0, not 1, because even such simple
45 numbers as 0.6 and 0.2 cannot be presented exactly by floating-point
46 numbers. What you think in the above as 'three' is really more like
47 2.9999999999999995559.
49 =head2 Why isn't my octal data interpreted correctly?
51 Perl only understands octal and hex numbers as such when they occur as
52 literals in your program. Octal literals in perl must start with a
53 leading C<0> and hexadecimal literals must start with a leading C<0x>.
54 If they are read in from somewhere and assigned, no automatic
55 conversion takes place. You must explicitly use C<oct()> or C<hex()> if you
56 want the values converted to decimal. C<oct()> interprets hexadecimal (C<0x350>),
57 octal (C<0350> or even without the leading C<0>, like C<377>) and binary
58 (C<0b1010>) numbers, while C<hex()> only converts hexadecimal ones, with
59 or without a leading C<0x>, such as C<0x255>, C<3A>, C<ff>, or C<deadbeef>.
60 The inverse mapping from decimal to octal can be done with either the
61 <%o> or C<%O> C<sprintf()> formats.
63 This problem shows up most often when people try using C<chmod()>,
64 C<mkdir()>, C<umask()>, or C<sysopen()>, which by widespread tradition
65 typically take permissions in octal.
67 chmod(644, $file); # WRONG
68 chmod(0644, $file); # right
70 Note the mistake in the first line was specifying the decimal literal
71 C<644>, rather than the intended octal literal C<0644>. The problem can
74 printf("%#o",644); # prints 01204
76 Surely you had not intended C<chmod(01204, $file);> - did you? If you
77 want to use numeric literals as arguments to chmod() et al. then please
78 try to express them as octal constants, that is with a leading zero and
79 with the following digits restricted to the set C<0..7>.
81 =head2 Does Perl have a round() function? What about ceil() and floor()? Trig functions?
83 Remember that C<int()> merely truncates toward 0. For rounding to a
84 certain number of digits, C<sprintf()> or C<printf()> is usually the
87 printf("%.3f", 3.1415926535); # prints 3.142
89 The C<POSIX> module (part of the standard Perl distribution)
90 implements C<ceil()>, C<floor()>, and a number of other mathematical
91 and trigonometric functions.
94 $ceil = ceil(3.5); # 4
95 $floor = floor(3.5); # 3
97 In 5.000 to 5.003 perls, trigonometry was done in the C<Math::Complex>
98 module. With 5.004, the C<Math::Trig> module (part of the standard Perl
99 distribution) implements the trigonometric functions. Internally it
100 uses the C<Math::Complex> module and some functions can break out from
101 the real axis into the complex plane, for example the inverse sine of
104 Rounding in financial applications can have serious implications, and
105 the rounding method used should be specified precisely. In these
106 cases, it probably pays not to trust whichever system rounding is
107 being used by Perl, but to instead implement the rounding function you
110 To see why, notice how you'll still have an issue on half-way-point
113 for ($i = 0; $i < 1.01; $i += 0.05) { printf "%.1f ",$i}
115 0.0 0.1 0.1 0.2 0.2 0.2 0.3 0.3 0.4 0.4 0.5 0.5 0.6 0.7 0.7
116 0.8 0.8 0.9 0.9 1.0 1.0
118 Don't blame Perl. It's the same as in C. IEEE says we have to do
119 this. Perl numbers whose absolute values are integers under 2**31 (on
120 32 bit machines) will work pretty much like mathematical integers.
121 Other numbers are not guaranteed.
123 =head2 How do I convert between numeric representations/bases/radixes?
125 As always with Perl there is more than one way to do it. Below are a
126 few examples of approaches to making common conversions between number
127 representations. This is intended to be representational rather than
130 Some of the examples later in L<perlfaq4> use the C<Bit::Vector>
131 module from CPAN. The reason you might choose C<Bit::Vector> over the
132 perl built in functions is that it works with numbers of ANY size,
133 that it is optimized for speed on some operations, and for at least
134 some programmers the notation might be familiar.
138 =item How do I convert hexadecimal into decimal
140 Using perl's built in conversion of C<0x> notation:
144 Using the C<hex> function:
146 $dec = hex("DEADBEEF");
150 $dec = unpack("N", pack("H8", substr("0" x 8 . "DEADBEEF", -8)));
152 Using the CPAN module C<Bit::Vector>:
155 $vec = Bit::Vector->new_Hex(32, "DEADBEEF");
156 $dec = $vec->to_Dec();
158 =item How do I convert from decimal to hexadecimal
162 $hex = sprintf("%X", 3735928559); # upper case A-F
163 $hex = sprintf("%x", 3735928559); # lower case a-f
167 $hex = unpack("H*", pack("N", 3735928559));
169 Using C<Bit::Vector>:
172 $vec = Bit::Vector->new_Dec(32, -559038737);
173 $hex = $vec->to_Hex();
175 And C<Bit::Vector> supports odd bit counts:
178 $vec = Bit::Vector->new_Dec(33, 3735928559);
179 $vec->Resize(32); # suppress leading 0 if unwanted
180 $hex = $vec->to_Hex();
182 =item How do I convert from octal to decimal
184 Using Perl's built in conversion of numbers with leading zeros:
186 $dec = 033653337357; # note the leading 0!
188 Using the C<oct> function:
190 $dec = oct("33653337357");
192 Using C<Bit::Vector>:
195 $vec = Bit::Vector->new(32);
196 $vec->Chunk_List_Store(3, split(//, reverse "33653337357"));
197 $dec = $vec->to_Dec();
199 =item How do I convert from decimal to octal
203 $oct = sprintf("%o", 3735928559);
205 Using C<Bit::Vector>:
208 $vec = Bit::Vector->new_Dec(32, -559038737);
209 $oct = reverse join('', $vec->Chunk_List_Read(3));
211 =item How do I convert from binary to decimal
213 Perl 5.6 lets you write binary numbers directly with
216 $number = 0b10110110;
220 my $input = "10110110";
221 $decimal = oct( "0b$input" );
223 Using C<pack> and C<ord>:
225 $decimal = ord(pack('B8', '10110110'));
227 Using C<pack> and C<unpack> for larger strings:
229 $int = unpack("N", pack("B32",
230 substr("0" x 32 . "11110101011011011111011101111", -32)));
231 $dec = sprintf("%d", $int);
233 # substr() is used to left pad a 32 character string with zeros.
235 Using C<Bit::Vector>:
237 $vec = Bit::Vector->new_Bin(32, "11011110101011011011111011101111");
238 $dec = $vec->to_Dec();
240 =item How do I convert from decimal to binary
242 Using C<sprintf> (perl 5.6+):
244 $bin = sprintf("%b", 3735928559);
248 $bin = unpack("B*", pack("N", 3735928559));
250 Using C<Bit::Vector>:
253 $vec = Bit::Vector->new_Dec(32, -559038737);
254 $bin = $vec->to_Bin();
256 The remaining transformations (e.g. hex -> oct, bin -> hex, etc.)
257 are left as an exercise to the inclined reader.
261 =head2 Why doesn't & work the way I want it to?
263 The behavior of binary arithmetic operators depends on whether they're
264 used on numbers or strings. The operators treat a string as a series
265 of bits and work with that (the string C<"3"> is the bit pattern
266 C<00110011>). The operators work with the binary form of a number
267 (the number C<3> is treated as the bit pattern C<00000011>).
269 So, saying C<11 & 3> performs the "and" operation on numbers (yielding
270 C<3>). Saying C<"11" & "3"> performs the "and" operation on strings
273 Most problems with C<&> and C<|> arise because the programmer thinks
274 they have a number but really it's a string. The rest arise because
277 if ("\020\020" & "\101\101") {
281 but a string consisting of two null bytes (the result of C<"\020\020"
282 & "\101\101">) is not a false value in Perl. You need:
284 if ( ("\020\020" & "\101\101") !~ /[^\000]/) {
288 =head2 How do I multiply matrices?
290 Use the Math::Matrix or Math::MatrixReal modules (available from CPAN)
291 or the PDL extension (also available from CPAN).
293 =head2 How do I perform an operation on a series of integers?
295 To call a function on each element in an array, and collect the
298 @results = map { my_func($_) } @array;
302 @triple = map { 3 * $_ } @single;
304 To call a function on each element of an array, but ignore the
307 foreach $iterator (@array) {
308 some_func($iterator);
311 To call a function on each integer in a (small) range, you B<can> use:
313 @results = map { some_func($_) } (5 .. 25);
315 but you should be aware that the C<..> operator creates an array of
316 all integers in the range. This can take a lot of memory for large
320 for ($i=5; $i < 500_005; $i++) {
321 push(@results, some_func($i));
324 This situation has been fixed in Perl5.005. Use of C<..> in a C<for>
325 loop will iterate over the range, without creating the entire range.
327 for my $i (5 .. 500_005) {
328 push(@results, some_func($i));
331 will not create a list of 500,000 integers.
333 =head2 How can I output Roman numerals?
335 Get the http://www.cpan.org/modules/by-module/Roman module.
337 =head2 Why aren't my random numbers random?
339 If you're using a version of Perl before 5.004, you must call C<srand>
340 once at the start of your program to seed the random number generator.
342 BEGIN { srand() if $] < 5.004 }
344 5.004 and later automatically call C<srand> at the beginning. Don't
345 call C<srand> more than once--you make your numbers less random,
348 Computers are good at being predictable and bad at being random
349 (despite appearances caused by bugs in your programs :-). see the
350 F<random> article in the "Far More Than You Ever Wanted To Know"
351 collection in http://www.cpan.org/misc/olddoc/FMTEYEWTK.tgz , courtesy
352 of Tom Phoenix, talks more about this. John von Neumann said, "Anyone
353 who attempts to generate random numbers by deterministic means is, of
354 course, living in a state of sin."
356 If you want numbers that are more random than C<rand> with C<srand>
357 provides, you should also check out the C<Math::TrulyRandom> module from
358 CPAN. It uses the imperfections in your system's timer to generate
359 random numbers, but this takes quite a while. If you want a better
360 pseudorandom generator than comes with your operating system, look at
361 "Numerical Recipes in C" at http://www.nr.com/ .
363 =head2 How do I get a random number between X and Y?
365 To get a random number between two values, you can use the C<rand()>
366 builtin to get a random number between 0 and 1. From there, you shift
367 that into the range that you want.
369 C<rand($x)> returns a number such that C<< 0 <= rand($x) < $x >>. Thus
370 what you want to have perl figure out is a random number in the range
371 from 0 to the difference between your I<X> and I<Y>.
373 That is, to get a number between 10 and 15, inclusive, you want a
374 random number between 0 and 5 that you can then add to 10.
376 my $number = 10 + int rand( 15-10+1 );
378 Hence you derive the following simple function to abstract
379 that. It selects a random integer between the two given
380 integers (inclusive), For example: C<random_int_between(50,120)>.
382 sub random_int_between {
384 # Assumes that the two arguments are integers themselves!
385 return $min if $min == $max;
386 ($min, $max) = ($max, $min) if $min > $max;
387 return $min + int rand(1 + $max - $min);
392 =head2 How do I find the day or week of the year?
394 The localtime function returns the day of the year. Without an
395 argument localtime uses the current time.
397 $day_of_year = (localtime)[7];
399 The C<POSIX> module can also format a date as the day of the year or
402 use POSIX qw/strftime/;
403 my $day_of_year = strftime "%j", localtime;
404 my $week_of_year = strftime "%W", localtime;
406 To get the day of year for any date, use C<POSIX>'s C<mktime> to get
407 a time in epoch seconds for the argument to localtime.
409 use POSIX qw/mktime strftime/;
410 my $week_of_year = strftime "%W",
411 localtime( mktime( 0, 0, 0, 18, 11, 87 ) );
413 The C<Date::Calc> module provides two functions to calculate these.
416 my $day_of_year = Day_of_Year( 1987, 12, 18 );
417 my $week_of_year = Week_of_Year( 1987, 12, 18 );
419 =head2 How do I find the current century or millennium?
421 Use the following simple functions:
424 return int((((localtime(shift || time))[5] + 1999))/100);
428 return 1+int((((localtime(shift || time))[5] + 1899))/1000);
431 On some systems, the C<POSIX> module's C<strftime()> function has been
432 extended in a non-standard way to use a C<%C> format, which they
433 sometimes claim is the "century". It isn't, because on most such
434 systems, this is only the first two digits of the four-digit year, and
435 thus cannot be used to reliably determine the current century or
438 =head2 How can I compare two dates and find the difference?
440 (contributed by brian d foy)
442 You could just store all your dates as a number and then subtract.
443 Life isn't always that simple though. If you want to work with
444 formatted dates, the C<Date::Manip>, C<Date::Calc>, or C<DateTime>
445 modules can help you.
447 =head2 How can I take a string and turn it into epoch seconds?
449 If it's a regular enough string that it always has the same format,
450 you can split it up and pass the parts to C<timelocal> in the standard
451 C<Time::Local> module. Otherwise, you should look into the C<Date::Calc>
452 and C<Date::Manip> modules from CPAN.
454 =head2 How can I find the Julian Day?
456 (contributed by brian d foy and Dave Cross)
458 You can use the C<Time::JulianDay> module available on CPAN. Ensure
459 that you really want to find a Julian day, though, as many people have
460 different ideas about Julian days. See
461 http://www.hermetic.ch/cal_stud/jdn.htm for instance.
463 You can also try the C<DateTime> module, which can convert a date/time
466 $ perl -MDateTime -le'print DateTime->today->jd'
469 Or the modified Julian Day
471 $ perl -MDateTime -le'print DateTime->today->mjd'
474 Or even the day of the year (which is what some people think of as a
477 $ perl -MDateTime -le'print DateTime->today->doy'
480 =head2 How do I find yesterday's date?
482 (contributed by brian d foy)
484 Use one of the Date modules. The C<DateTime> module makes it simple, and
485 give you the same time of day, only the day before.
489 my $yesterday = DateTime->now->subtract( days => 1 );
491 print "Yesterday was $yesterday\n";
493 You can also use the C<Date::Calc> module using its C<Today_and_Now>
496 use Date::Calc qw( Today_and_Now Add_Delta_DHMS );
498 my @date_time = Add_Delta_DHMS( Today_and_Now(), -1, 0, 0, 0 );
500 print "@date_time\n";
502 Most people try to use the time rather than the calendar to figure out
503 dates, but that assumes that days are twenty-four hours each. For
504 most people, there are two days a year when they aren't: the switch to
505 and from summer time throws this off. Let the modules do the work.
507 =head2 Does Perl have a Year 2000 problem? Is Perl Y2K compliant?
509 Short answer: No, Perl does not have a Year 2000 problem. Yes, Perl is
510 Y2K compliant (whatever that means). The programmers you've hired to
511 use it, however, probably are not.
513 Long answer: The question belies a true understanding of the issue.
514 Perl is just as Y2K compliant as your pencil--no more, and no less.
515 Can you use your pencil to write a non-Y2K-compliant memo? Of course
516 you can. Is that the pencil's fault? Of course it isn't.
518 The date and time functions supplied with Perl (gmtime and localtime)
519 supply adequate information to determine the year well beyond 2000
520 (2038 is when trouble strikes for 32-bit machines). The year returned
521 by these functions when used in a list context is the year minus 1900.
522 For years between 1910 and 1999 this I<happens> to be a 2-digit decimal
523 number. To avoid the year 2000 problem simply do not treat the year as
524 a 2-digit number. It isn't.
526 When gmtime() and localtime() are used in scalar context they return
527 a timestamp string that contains a fully-expanded year. For example,
528 C<$timestamp = gmtime(1005613200)> sets $timestamp to "Tue Nov 13 01:00:00
529 2001". There's no year 2000 problem here.
531 That doesn't mean that Perl can't be used to create non-Y2K compliant
532 programs. It can. But so can your pencil. It's the fault of the user,
533 not the language. At the risk of inflaming the NRA: "Perl doesn't
534 break Y2K, people do." See http://www.perl.org/about/y2k.html for
539 =head2 How do I validate input?
541 (contributed by brian d foy)
543 There are many ways to ensure that values are what you expect or
544 want to accept. Besides the specific examples that we cover in the
545 perlfaq, you can also look at the modules with "Assert" and "Validate"
546 in their names, along with other modules such as C<Regexp::Common>.
548 Some modules have validation for particular types of input, such
549 as C<Business::ISBN>, C<Business::CreditCard>, C<Email::Valid>,
550 and C<Data::Validate::IP>.
552 =head2 How do I unescape a string?
554 It depends just what you mean by "escape". URL escapes are dealt
555 with in L<perlfaq9>. Shell escapes with the backslash (C<\>)
556 character are removed with
560 This won't expand C<"\n"> or C<"\t"> or any other special escapes.
562 =head2 How do I remove consecutive pairs of characters?
564 (contributed by brian d foy)
566 You can use the substitution operator to find pairs of characters (or
567 runs of characters) and replace them with a single instance. In this
568 substitution, we find a character in C<(.)>. The memory parentheses
569 store the matched character in the back-reference C<\1> and we use
570 that to require that the same thing immediately follow it. We replace
571 that part of the string with the character in C<$1>.
575 We can also use the transliteration operator, C<tr///>. In this
576 example, the search list side of our C<tr///> contains nothing, but
577 the C<c> option complements that so it contains everything. The
578 replacement list also contains nothing, so the transliteration is
579 almost a no-op since it won't do any replacements (or more exactly,
580 replace the character with itself). However, the C<s> option squashes
581 duplicated and consecutive characters in the string so a character
582 does not show up next to itself
584 my $str = 'Haarlem'; # in the Netherlands
585 $str =~ tr///cs; # Now Harlem, like in New York
587 =head2 How do I expand function calls in a string?
589 (contributed by brian d foy)
591 This is documented in L<perlref>, and although it's not the easiest
592 thing to read, it does work. In each of these examples, we call the
593 function inside the braces used to dereference a reference. If we
594 have more than one return value, we can construct and dereference an
595 anonymous array. In this case, we call the function in list context.
597 print "The time values are @{ [localtime] }.\n";
599 If we want to call the function in scalar context, we have to do a bit
600 more work. We can really have any code we like inside the braces, so
601 we simply have to end with the scalar reference, although how you do
602 that is up to you, and you can use code inside the braces. Note that
603 the use of parens creates a list context, so we need C<scalar> to
604 force the scalar context on the function:
606 print "The time is ${\(scalar localtime)}.\n"
608 print "The time is ${ my $x = localtime; \$x }.\n";
610 If your function already returns a reference, you don't need to create
611 the reference yourself.
613 sub timestamp { my $t = localtime; \$t }
615 print "The time is ${ timestamp() }.\n";
617 The C<Interpolation> module can also do a lot of magic for you. You can
618 specify a variable name, in this case C<E>, to set up a tied hash that
619 does the interpolation for you. It has several other methods to do this
622 use Interpolation E => 'eval';
623 print "The time values are $E{localtime()}.\n";
625 In most cases, it is probably easier to simply use string concatenation,
626 which also forces scalar context.
628 print "The time is " . localtime() . ".\n";
630 =head2 How do I find matching/nesting anything?
632 This isn't something that can be done in one regular expression, no
633 matter how complicated. To find something between two single
634 characters, a pattern like C</x([^x]*)x/> will get the intervening
635 bits in $1. For multiple ones, then something more like
636 C</alpha(.*?)omega/> would be needed. But none of these deals with
637 nested patterns. For balanced expressions using C<(>, C<{>, C<[> or
638 C<< < >> as delimiters, use the CPAN module Regexp::Common, or see
639 L<perlre/(??{ code })>. For other cases, you'll have to write a
642 If you are serious about writing a parser, there are a number of
643 modules or oddities that will make your life a lot easier. There are
644 the CPAN modules C<Parse::RecDescent>, C<Parse::Yapp>, and
645 C<Text::Balanced>; and the C<byacc> program. Starting from perl 5.8
646 the C<Text::Balanced> is part of the standard distribution.
648 One simple destructive, inside-out approach that you might try is to
649 pull out the smallest nesting parts one at a time:
651 while (s/BEGIN((?:(?!BEGIN)(?!END).)*)END//gs) {
652 # do something with $1
655 A more complicated and sneaky approach is to make Perl's regular
656 expression engine do it for you. This is courtesy Dean Inada, and
657 rather has the nature of an Obfuscated Perl Contest entry, but it
660 # $_ contains the string to parse
661 # BEGIN and END are the opening and closing markers for the
666 ($re=$_)=~s/((BEGIN)|(END)|.)/$)[!$3]\Q$1\E$([!$2]/gs;
667 @$ = (eval{/$re/},$@!~/unmatched/i);
668 print join("\n",@$[0..$#$]) if( $$[-1] );
670 =head2 How do I reverse a string?
672 Use C<reverse()> in scalar context, as documented in
675 $reversed = reverse $string;
677 =head2 How do I expand tabs in a string?
679 You can do it yourself:
681 1 while $string =~ s/\t+/' ' x (length($&) * 8 - length($`) % 8)/e;
683 Or you can just use the C<Text::Tabs> module (part of the standard Perl
687 @expanded_lines = expand(@lines_with_tabs);
689 =head2 How do I reformat a paragraph?
691 Use C<Text::Wrap> (part of the standard Perl distribution):
694 print wrap("\t", ' ', @paragraphs);
696 The paragraphs you give to C<Text::Wrap> should not contain embedded
697 newlines. C<Text::Wrap> doesn't justify the lines (flush-right).
699 Or use the CPAN module C<Text::Autoformat>. Formatting files can be
700 easily done by making a shell alias, like so:
702 alias fmt="perl -i -MText::Autoformat -n0777 \
703 -e 'print autoformat $_, {all=>1}' $*"
705 See the documentation for C<Text::Autoformat> to appreciate its many
708 =head2 How can I access or change N characters of a string?
710 You can access the first characters of a string with substr().
711 To get the first character, for example, start at position 0
712 and grab the string of length 1.
715 $string = "Just another Perl Hacker";
716 $first_char = substr( $string, 0, 1 ); # 'J'
718 To change part of a string, you can use the optional fourth
719 argument which is the replacement string.
721 substr( $string, 13, 4, "Perl 5.8.0" );
723 You can also use substr() as an lvalue.
725 substr( $string, 13, 4 ) = "Perl 5.8.0";
727 =head2 How do I change the Nth occurrence of something?
729 You have to keep track of N yourself. For example, let's say you want
730 to change the fifth occurrence of C<"whoever"> or C<"whomever"> into
731 C<"whosoever"> or C<"whomsoever">, case insensitively. These
732 all assume that $_ contains the string to be altered.
736 ++$count == 5 # is it the 5th?
737 ? "${2}soever" # yes, swap
738 : $1 # renege and leave it there
741 In the more general case, you can use the C</g> modifier in a C<while>
742 loop, keeping count of matches.
746 $_ = "One fish two fish red fish blue fish";
747 while (/(\w+)\s+fish\b/gi) {
748 if (++$count == $WANT) {
749 print "The third fish is a $1 one.\n";
753 That prints out: C<"The third fish is a red one."> You can also use a
754 repetition count and repeated pattern like this:
756 /(?:\w+\s+fish\s+){2}(\w+)\s+fish/i;
758 =head2 How can I count the number of occurrences of a substring within a string?
760 There are a number of ways, with varying efficiency. If you want a
761 count of a certain single character (X) within a string, you can use the
762 C<tr///> function like so:
764 $string = "ThisXlineXhasXsomeXx'sXinXit";
765 $count = ($string =~ tr/X//);
766 print "There are $count X characters in the string";
768 This is fine if you are just looking for a single character. However,
769 if you are trying to count multiple character substrings within a
770 larger string, C<tr///> won't work. What you can do is wrap a while()
771 loop around a global pattern match. For example, let's count negative
774 $string = "-9 55 48 -2 23 -76 4 14 -44";
775 while ($string =~ /-\d+/g) { $count++ }
776 print "There are $count negative numbers in the string";
778 Another version uses a global match in list context, then assigns the
779 result to a scalar, producing a count of the number of matches.
781 $count = () = $string =~ /-\d+/g;
783 =head2 How do I capitalize all the words on one line?
785 To make the first letter of each word upper case:
787 $line =~ s/\b(\w)/\U$1/g;
789 This has the strange effect of turning "C<don't do it>" into "C<Don'T
790 Do It>". Sometimes you might want this. Other times you might need a
791 more thorough solution (Suggested by brian d foy):
794 (^\w) #at the beginning of the line
796 (\s\w) #preceded by whitespace
800 $string =~ s/([\w']+)/\u\L$1/g;
802 To make the whole line upper case:
806 To force each word to be lower case, with the first letter upper case:
808 $line =~ s/(\w+)/\u\L$1/g;
810 You can (and probably should) enable locale awareness of those
811 characters by placing a C<use locale> pragma in your program.
812 See L<perllocale> for endless details on locales.
814 This is sometimes referred to as putting something into "title
815 case", but that's not quite accurate. Consider the proper
816 capitalization of the movie I<Dr. Strangelove or: How I Learned to
817 Stop Worrying and Love the Bomb>, for example.
819 Damian Conway's L<Text::Autoformat> module provides some smart
820 case transformations:
822 use Text::Autoformat;
823 my $x = "Dr. Strangelove or: How I Learned to Stop ".
824 "Worrying and Love the Bomb";
827 for my $style (qw( sentence title highlight )) {
828 print autoformat($x, { case => $style }), "\n";
831 =head2 How can I split a [character] delimited string except when inside [character]?
833 Several modules can handle this sort of parsing--C<Text::Balanced>,
834 C<Text::CSV>, C<Text::CSV_XS>, and C<Text::ParseWords>, among others.
836 Take the example case of trying to split a string that is
837 comma-separated into its different fields. You can't use C<split(/,/)>
838 because you shouldn't split if the comma is inside quotes. For
839 example, take a data line like this:
841 SAR001,"","Cimetrix, Inc","Bob Smith","CAM",N,8,1,0,7,"Error, Core Dumped"
843 Due to the restriction of the quotes, this is a fairly complex
844 problem. Thankfully, we have Jeffrey Friedl, author of
845 I<Mastering Regular Expressions>, to handle these for us. He
846 suggests (assuming your string is contained in C<$text>):
849 push(@new, $+) while $text =~ m{
850 "([^\"\\]*(?:\\.[^\"\\]*)*)",? # groups the phrase inside the quotes
854 push(@new, undef) if substr($text,-1,1) eq ',';
856 If you want to represent quotation marks inside a
857 quotation-mark-delimited field, escape them with backslashes (eg,
860 Alternatively, the C<Text::ParseWords> module (part of the standard
861 Perl distribution) lets you say:
863 use Text::ParseWords;
864 @new = quotewords(",", 0, $text);
866 =head2 How do I strip blank space from the beginning/end of a string?
868 (contributed by brian d foy)
870 A substitution can do this for you. For a single line, you want to
871 replace all the leading or trailing whitespace with nothing. You
872 can do that with a pair of substitutions.
877 You can also write that as a single substitution, although it turns
878 out the combined statement is slower than the separate ones. That
879 might not matter to you, though.
883 In this regular expression, the alternation matches either at the
884 beginning or the end of the string since the anchors have a lower
885 precedence than the alternation. With the C</g> flag, the substitution
886 makes all possible matches, so it gets both. Remember, the trailing
887 newline matches the C<\s+>, and the C<$> anchor can match to the
888 physical end of the string, so the newline disappears too. Just add
889 the newline to the output, which has the added benefit of preserving
890 "blank" (consisting entirely of whitespace) lines which the C<^\s+>
891 would remove all by itself.
899 For a multi-line string, you can apply the regular expression
900 to each logical line in the string by adding the C</m> flag (for
901 "multi-line"). With the C</m> flag, the C<$> matches I<before> an
902 embedded newline, so it doesn't remove it. It still removes the
903 newline at the end of the string.
905 $string =~ s/^\s+|\s+$//gm;
907 Remember that lines consisting entirely of whitespace will disappear,
908 since the first part of the alternation can match the entire string
909 and replace it with nothing. If need to keep embedded blank lines,
910 you have to do a little more work. Instead of matching any whitespace
911 (since that includes a newline), just match the other whitespace.
913 $string =~ s/^[\t\f ]+|[\t\f ]+$//mg;
915 =head2 How do I pad a string with blanks or pad a number with zeroes?
917 In the following examples, C<$pad_len> is the length to which you wish
918 to pad the string, C<$text> or C<$num> contains the string to be padded,
919 and C<$pad_char> contains the padding character. You can use a single
920 character string constant instead of the C<$pad_char> variable if you
921 know what it is in advance. And in the same way you can use an integer in
922 place of C<$pad_len> if you know the pad length in advance.
924 The simplest method uses the C<sprintf> function. It can pad on the left
925 or right with blanks and on the left with zeroes and it will not
926 truncate the result. The C<pack> function can only pad strings on the
927 right with blanks and it will truncate the result to a maximum length of
930 # Left padding a string with blanks (no truncation):
931 $padded = sprintf("%${pad_len}s", $text);
932 $padded = sprintf("%*s", $pad_len, $text); # same thing
934 # Right padding a string with blanks (no truncation):
935 $padded = sprintf("%-${pad_len}s", $text);
936 $padded = sprintf("%-*s", $pad_len, $text); # same thing
938 # Left padding a number with 0 (no truncation):
939 $padded = sprintf("%0${pad_len}d", $num);
940 $padded = sprintf("%0*d", $pad_len, $num); # same thing
942 # Right padding a string with blanks using pack (will truncate):
943 $padded = pack("A$pad_len",$text);
945 If you need to pad with a character other than blank or zero you can use
946 one of the following methods. They all generate a pad string with the
947 C<x> operator and combine that with C<$text>. These methods do
948 not truncate C<$text>.
950 Left and right padding with any character, creating a new string:
952 $padded = $pad_char x ( $pad_len - length( $text ) ) . $text;
953 $padded = $text . $pad_char x ( $pad_len - length( $text ) );
955 Left and right padding with any character, modifying C<$text> directly:
957 substr( $text, 0, 0 ) = $pad_char x ( $pad_len - length( $text ) );
958 $text .= $pad_char x ( $pad_len - length( $text ) );
960 =head2 How do I extract selected columns from a string?
962 (contributed by brian d foy)
964 If you know where the columns that contain the data, you can
965 use C<substr> to extract a single column.
967 my $column = substr( $line, $start_column, $length );
969 You can use C<split> if the columns are separated by whitespace or
970 some other delimiter, as long as whitespace or the delimiter cannot
971 appear as part of the data.
973 my $line = ' fred barney betty ';
974 my @columns = split /\s+/, $line;
975 # ( '', 'fred', 'barney', 'betty' );
977 my $line = 'fred||barney||betty';
978 my @columns = split /\|/, $line;
979 # ( 'fred', '', 'barney', '', 'betty' );
981 If you want to work with comma-separated values, don't do this since
982 that format is a bit more complicated. Use one of the modules that
983 handle that fornat, such as C<Text::CSV>, C<Text::CSV_XS>, or
986 If you want to break apart an entire line of fixed columns, you can use
987 C<unpack> with the A (ASCII) format. by using a number after the format
988 specifier, you can denote the column width. See the C<pack> and C<unpack>
989 entries in L<perlfunc> for more details.
991 my @fields = unpack( $line, "A8 A8 A8 A16 A4" );
993 Note that spaces in the format argument to C<unpack> do not denote literal
994 spaces. If you have space separated data, you may want C<split> instead.
996 =head2 How do I find the soundex value of a string?
998 (contributed by brian d foy)
1000 You can use the Text::Soundex module. If you want to do fuzzy or close
1001 matching, you might also try the C<String::Approx>, and
1002 C<Text::Metaphone>, and C<Text::DoubleMetaphone> modules.
1004 =head2 How can I expand variables in text strings?
1006 (contributed by brian d foy)
1008 If you can avoid it, don't, or if you can use a templating system,
1009 such as C<Text::Template> or C<Template> Toolkit, do that instead.
1011 However, for the one-off simple case where I don't want to pull out a
1012 full templating system, I'll use a string that has two Perl scalar
1013 variables in it. In this example, I want to expand C<$foo> and C<$bar>
1014 to their variable's values.
1018 $string = 'Say hello to $foo and $bar';
1020 One way I can do this involves the substitution operator and a double
1021 C</e> flag. The first C</e> evaluates C<$1> on the replacement side and
1022 turns it into C<$foo>. The second /e starts with C<$foo> and replaces
1023 it with its value. C<$foo>, then, turns into 'Fred', and that's finally
1024 what's left in the string.
1026 $string =~ s/(\$\w+)/$1/eeg; # 'Say hello to Fred and Barney'
1028 The C</e> will also silently ignore violations of strict, replacing
1029 undefined variable names with the empty string.
1031 I could also pull the values from a hash instead of evaluating
1032 variable names. Using a single C</e>, I can check the hash to ensure
1033 the value exists, and if it doesn't, I can replace the missing value
1034 with a marker, in this case C<???> to signal that I missed something:
1036 my $string = 'This has $foo and $bar';
1038 my %Replacements = (
1042 # $string =~ s/\$(\w+)/$Replacements{$1}/g;
1043 $string =~ s/\$(\w+)/
1044 exists $Replacements{$1} ? $Replacements{$1} : '???'
1049 =head2 What's wrong with always quoting "$vars"?
1051 The problem is that those double-quotes force
1052 stringification--coercing numbers and references into strings--even
1053 when you don't want them to be strings. Think of it this way:
1054 double-quote expansion is used to produce new strings. If you already
1055 have a string, why do you need more?
1057 If you get used to writing odd things like these:
1060 $new = "$old"; # BAD
1061 somefunc("$var"); # BAD
1063 You'll be in trouble. Those should (in 99.8% of the cases) be
1064 the simpler and more direct:
1070 Otherwise, besides slowing you down, you're going to break code when
1071 the thing in the scalar is actually neither a string nor a number, but
1077 my $oref = "$aref"; # WRONG
1080 You can also get into subtle problems on those few operations in Perl
1081 that actually do care about the difference between a string and a
1082 number, such as the magical C<++> autoincrement operator or the
1085 Stringification also destroys arrays.
1088 print "@lines"; # WRONG - extra blanks
1089 print @lines; # right
1091 =head2 Why don't my E<lt>E<lt>HERE documents work?
1093 Check for these three things:
1097 =item There must be no space after the E<lt>E<lt> part.
1099 =item There (probably) should be a semicolon at the end.
1101 =item You can't (easily) have any space in front of the tag.
1105 If you want to indent the text in the here document, you
1109 ($VAR = <<HERE_TARGET) =~ s/^\s+//gm;
1114 But the HERE_TARGET must still be flush against the margin.
1115 If you want that indented also, you'll have to quote
1118 ($quote = <<' FINIS') =~ s/^\s+//gm;
1119 ...we will have peace, when you and all your works have
1120 perished--and the works of your dark master to whom you
1121 would deliver us. You are a liar, Saruman, and a corrupter
1122 of men's hearts. --Theoden in /usr/src/perl/taint.c
1124 $quote =~ s/\s+--/\n--/;
1126 A nice general-purpose fixer-upper function for indented here documents
1127 follows. It expects to be called with a here document as its argument.
1128 It looks to see whether each line begins with a common substring, and
1129 if so, strips that substring off. Otherwise, it takes the amount of leading
1130 whitespace found on the first line and removes that much off each
1135 my ($white, $leader); # common whitespace and common leading string
1136 if (/^\s*(?:([^\w\s]+)(\s*).*\n)(?:\s*\1\2?.*\n)+$/) {
1137 ($white, $leader) = ($2, quotemeta($1));
1139 ($white, $leader) = (/^(\s+)/, '');
1141 s/^\s*?$leader(?:$white)?//gm;
1145 This works with leading special strings, dynamically determined:
1147 $remember_the_main = fix<<' MAIN_INTERPRETER_LOOP';
1150 @@@ SAVEI32(runlevel);
1152 @@@ while ( op = (*op->op_ppaddr)() );
1156 MAIN_INTERPRETER_LOOP
1158 Or with a fixed amount of leading whitespace, with remaining
1159 indentation correctly preserved:
1161 $poem = fix<<EVER_ON_AND_ON;
1162 Now far ahead the Road has gone,
1163 And I must follow, if I can,
1164 Pursuing it with eager feet,
1165 Until it joins some larger way
1166 Where many paths and errands meet.
1167 And whither then? I cannot say.
1168 --Bilbo in /usr/src/perl/pp_ctl.c
1173 =head2 What is the difference between a list and an array?
1175 An array has a changeable length. A list does not. An array is
1176 something you can push or pop, while a list is a set of values. Some
1177 people make the distinction that a list is a value while an array is a
1178 variable. Subroutines are passed and return lists, you put things into
1179 list context, you initialize arrays with lists, and you C<foreach()>
1180 across a list. C<@> variables are arrays, anonymous arrays are
1181 arrays, arrays in scalar context behave like the number of elements in
1182 them, subroutines access their arguments through the array C<@_>, and
1183 C<push>/C<pop>/C<shift> only work on arrays.
1185 As a side note, there's no such thing as a list in scalar context.
1188 $scalar = (2, 5, 7, 9);
1190 you're using the comma operator in scalar context, so it uses the scalar
1191 comma operator. There never was a list there at all! This causes the
1192 last value to be returned: 9.
1194 =head2 What is the difference between $array[1] and @array[1]?
1196 The former is a scalar value; the latter an array slice, making
1197 it a list with one (scalar) value. You should use $ when you want a
1198 scalar value (most of the time) and @ when you want a list with one
1199 scalar value in it (very, very rarely; nearly never, in fact).
1201 Sometimes it doesn't make a difference, but sometimes it does.
1202 For example, compare:
1204 $good[0] = `some program that outputs several lines`;
1208 @bad[0] = `same program that outputs several lines`;
1210 The C<use warnings> pragma and the B<-w> flag will warn you about these
1213 =head2 How can I remove duplicate elements from a list or array?
1215 (contributed by brian d foy)
1217 Use a hash. When you think the words "unique" or "duplicated", think
1220 If you don't care about the order of the elements, you could just
1221 create the hash then extract the keys. It's not important how you
1222 create that hash: just that you use C<keys> to get the unique
1225 my %hash = map { $_, 1 } @array;
1226 # or a hash slice: @hash{ @array } = ();
1227 # or a foreach: $hash{$_} = 1 foreach ( @array );
1229 my @unique = keys %hash;
1231 If you want to use a module, try the C<uniq> function from
1232 C<List::MoreUtils>. In list context it returns the unique elements,
1233 preserving their order in the list. In scalar context, it returns the
1234 number of unique elements.
1236 use List::MoreUtils qw(uniq);
1238 my @unique = uniq( 1, 2, 3, 4, 4, 5, 6, 5, 7 ); # 1,2,3,4,5,6,7
1239 my $unique = uniq( 1, 2, 3, 4, 4, 5, 6, 5, 7 ); # 7
1241 You can also go through each element and skip the ones you've seen
1242 before. Use a hash to keep track. The first time the loop sees an
1243 element, that element has no key in C<%Seen>. The C<next> statement
1244 creates the key and immediately uses its value, which is C<undef>, so
1245 the loop continues to the C<push> and increments the value for that
1246 key. The next time the loop sees that same element, its key exists in
1247 the hash I<and> the value for that key is true (since it's not 0 or
1248 C<undef>), so the next skips that iteration and the loop goes to the
1254 foreach my $elem ( @array )
1256 next if $seen{ $elem }++;
1257 push @unique, $elem;
1260 You can write this more briefly using a grep, which does the
1264 my @unique = grep { ! $seen{ $_ }++ } @array;
1266 =head2 How can I tell whether a certain element is contained in a list or array?
1268 (portions of this answer contributed by Anno Siegel)
1270 Hearing the word "in" is an I<in>dication that you probably should have
1271 used a hash, not a list or array, to store your data. Hashes are
1272 designed to answer this question quickly and efficiently. Arrays aren't.
1274 That being said, there are several ways to approach this. If you
1275 are going to make this query many times over arbitrary string values,
1276 the fastest way is probably to invert the original array and maintain a
1277 hash whose keys are the first array's values.
1279 @blues = qw/azure cerulean teal turquoise lapis-lazuli/;
1281 for (@blues) { $is_blue{$_} = 1 }
1283 Now you can check whether C<$is_blue{$some_color}>. It might have
1284 been a good idea to keep the blues all in a hash in the first place.
1286 If the values are all small integers, you could use a simple indexed
1287 array. This kind of an array will take up less space:
1289 @primes = (2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31);
1290 @is_tiny_prime = ();
1291 for (@primes) { $is_tiny_prime[$_] = 1 }
1292 # or simply @istiny_prime[@primes] = (1) x @primes;
1294 Now you check whether $is_tiny_prime[$some_number].
1296 If the values in question are integers instead of strings, you can save
1297 quite a lot of space by using bit strings instead:
1299 @articles = ( 1..10, 150..2000, 2017 );
1301 for (@articles) { vec($read,$_,1) = 1 }
1303 Now check whether C<vec($read,$n,1)> is true for some C<$n>.
1305 These methods guarantee fast individual tests but require a re-organization
1306 of the original list or array. They only pay off if you have to test
1307 multiple values against the same array.
1309 If you are testing only once, the standard module C<List::Util> exports
1310 the function C<first> for this purpose. It works by stopping once it
1311 finds the element. It's written in C for speed, and its Perl equivalant
1312 looks like this subroutine:
1317 return $_ if &{$code}();
1322 If speed is of little concern, the common idiom uses grep in scalar context
1323 (which returns the number of items that passed its condition) to traverse the
1324 entire list. This does have the benefit of telling you how many matches it
1327 my $is_there = grep $_ eq $whatever, @array;
1329 If you want to actually extract the matching elements, simply use grep in
1332 my @matches = grep $_ eq $whatever, @array;
1334 =head2 How do I compute the difference of two arrays? How do I compute the intersection of two arrays?
1336 Use a hash. Here's code to do both and more. It assumes that each
1337 element is unique in a given array:
1339 @union = @intersection = @difference = ();
1341 foreach $element (@array1, @array2) { $count{$element}++ }
1342 foreach $element (keys %count) {
1343 push @union, $element;
1344 push @{ $count{$element} > 1 ? \@intersection : \@difference }, $element;
1347 Note that this is the I<symmetric difference>, that is, all elements
1348 in either A or in B but not in both. Think of it as an xor operation.
1350 =head2 How do I test whether two arrays or hashes are equal?
1352 The following code works for single-level arrays. It uses a
1353 stringwise comparison, and does not distinguish defined versus
1354 undefined empty strings. Modify if you have other needs.
1356 $are_equal = compare_arrays(\@frogs, \@toads);
1358 sub compare_arrays {
1359 my ($first, $second) = @_;
1360 no warnings; # silence spurious -w undef complaints
1361 return 0 unless @$first == @$second;
1362 for (my $i = 0; $i < @$first; $i++) {
1363 return 0 if $first->[$i] ne $second->[$i];
1368 For multilevel structures, you may wish to use an approach more
1369 like this one. It uses the CPAN module C<FreezeThaw>:
1371 use FreezeThaw qw(cmpStr);
1372 @a = @b = ( "this", "that", [ "more", "stuff" ] );
1374 printf "a and b contain %s arrays\n",
1375 cmpStr(\@a, \@b) == 0
1379 This approach also works for comparing hashes. Here we'll demonstrate
1380 two different answers:
1382 use FreezeThaw qw(cmpStr cmpStrHard);
1384 %a = %b = ( "this" => "that", "extra" => [ "more", "stuff" ] );
1388 printf "a and b contain %s hashes\n",
1389 cmpStr(\%a, \%b) == 0 ? "the same" : "different";
1391 printf "a and b contain %s hashes\n",
1392 cmpStrHard(\%a, \%b) == 0 ? "the same" : "different";
1395 The first reports that both those the hashes contain the same data,
1396 while the second reports that they do not. Which you prefer is left as
1397 an exercise to the reader.
1399 =head2 How do I find the first array element for which a condition is true?
1401 To find the first array element which satisfies a condition, you can
1402 use the C<first()> function in the C<List::Util> module, which comes
1403 with Perl 5.8. This example finds the first element that contains
1406 use List::Util qw(first);
1408 my $element = first { /Perl/ } @array;
1410 If you cannot use C<List::Util>, you can make your own loop to do the
1411 same thing. Once you find the element, you stop the loop with last.
1414 foreach ( @array ) {
1415 if( /Perl/ ) { $found = $_; last }
1418 If you want the array index, you can iterate through the indices
1419 and check the array element at each index until you find one
1420 that satisfies the condition.
1422 my( $found, $index ) = ( undef, -1 );
1423 for( $i = 0; $i < @array; $i++ ) {
1424 if( $array[$i] =~ /Perl/ ) {
1425 $found = $array[$i];
1431 =head2 How do I handle linked lists?
1433 In general, you usually don't need a linked list in Perl, since with
1434 regular arrays, you can push and pop or shift and unshift at either
1435 end, or you can use splice to add and/or remove arbitrary number of
1436 elements at arbitrary points. Both pop and shift are both O(1)
1437 operations on Perl's dynamic arrays. In the absence of shifts and
1438 pops, push in general needs to reallocate on the order every log(N)
1439 times, and unshift will need to copy pointers each time.
1441 If you really, really wanted, you could use structures as described in
1442 L<perldsc> or L<perltoot> and do just what the algorithm book tells
1443 you to do. For example, imagine a list node like this:
1450 You could walk the list this way:
1453 for ($node = $head; $node; $node = $node->{LINK}) {
1454 print $node->{VALUE}, " ";
1458 You could add to the list this way:
1461 $tail = append($head, 1); # grow a new head
1462 for $value ( 2 .. 10 ) {
1463 $tail = append($tail, $value);
1467 my($list, $value) = @_;
1468 my $node = { VALUE => $value };
1470 $node->{LINK} = $list->{LINK};
1471 $list->{LINK} = $node;
1474 $_[0] = $node; # replace caller's version
1479 But again, Perl's built-in are virtually always good enough.
1481 =head2 How do I handle circular lists?
1483 Circular lists could be handled in the traditional fashion with linked
1484 lists, or you could just do something like this with an array:
1486 unshift(@array, pop(@array)); # the last shall be first
1487 push(@array, shift(@array)); # and vice versa
1489 You can also use C<Tie::Cycle>:
1493 tie my $cycle, 'Tie::Cycle', [ qw( FFFFFF 000000 FFFF00 ) ];
1495 print $cycle; # FFFFFF
1496 print $cycle; # 000000
1497 print $cycle; # FFFF00
1499 =head2 How do I shuffle an array randomly?
1501 If you either have Perl 5.8.0 or later installed, or if you have
1502 Scalar-List-Utils 1.03 or later installed, you can say:
1504 use List::Util 'shuffle';
1506 @shuffled = shuffle(@list);
1508 If not, you can use a Fisher-Yates shuffle.
1510 sub fisher_yates_shuffle {
1511 my $deck = shift; # $deck is a reference to an array
1514 my $j = int rand ($i+1);
1515 @$deck[$i,$j] = @$deck[$j,$i];
1519 # shuffle my mpeg collection
1521 my @mpeg = <audio/*/*.mp3>;
1522 fisher_yates_shuffle( \@mpeg ); # randomize @mpeg in place
1525 Note that the above implementation shuffles an array in place,
1526 unlike the C<List::Util::shuffle()> which takes a list and returns
1527 a new shuffled list.
1529 You've probably seen shuffling algorithms that work using splice,
1530 randomly picking another element to swap the current element with
1534 @old = 1 .. 10; # just a demo
1536 push(@new, splice(@old, rand @old, 1));
1539 This is bad because splice is already O(N), and since you do it N
1540 times, you just invented a quadratic algorithm; that is, O(N**2).
1541 This does not scale, although Perl is so efficient that you probably
1542 won't notice this until you have rather largish arrays.
1544 =head2 How do I process/modify each element of an array?
1546 Use C<for>/C<foreach>:
1549 s/foo/bar/; # change that word
1550 tr/XZ/ZX/; # swap those letters
1553 Here's another; let's compute spherical volumes:
1555 for (@volumes = @radii) { # @volumes has changed parts
1557 $_ *= (4/3) * 3.14159; # this will be constant folded
1560 which can also be done with C<map()> which is made to transform
1561 one list into another:
1563 @volumes = map {$_ ** 3 * (4/3) * 3.14159} @radii;
1565 If you want to do the same thing to modify the values of the
1566 hash, you can use the C<values> function. As of Perl 5.6
1567 the values are not copied, so if you modify $orbit (in this
1568 case), you modify the value.
1570 for $orbit ( values %orbits ) {
1571 ($orbit **= 3) *= (4/3) * 3.14159;
1574 Prior to perl 5.6 C<values> returned copies of the values,
1575 so older perl code often contains constructions such as
1576 C<@orbits{keys %orbits}> instead of C<values %orbits> where
1577 the hash is to be modified.
1579 =head2 How do I select a random element from an array?
1581 Use the C<rand()> function (see L<perlfunc/rand>):
1583 $index = rand @array;
1584 $element = $array[$index];
1588 my $element = $array[ rand @array ];
1590 =head2 How do I permute N elements of a list?
1592 Use the C<List::Permutor> module on CPAN. If the list is actually an
1593 array, try the C<Algorithm::Permute> module (also on CPAN). It's
1594 written in XS code and is very efficient.
1596 use Algorithm::Permute;
1597 my @array = 'a'..'d';
1598 my $p_iterator = Algorithm::Permute->new ( \@array );
1599 while (my @perm = $p_iterator->next) {
1600 print "next permutation: (@perm)\n";
1603 For even faster execution, you could do:
1605 use Algorithm::Permute;
1606 my @array = 'a'..'d';
1607 Algorithm::Permute::permute {
1608 print "next permutation: (@array)\n";
1611 Here's a little program that generates all permutations of
1612 all the words on each line of input. The algorithm embodied
1613 in the C<permute()> function is discussed in Volume 4 (still
1614 unpublished) of Knuth's I<The Art of Computer Programming>
1615 and will work on any list:
1618 # Fischer-Kause ordered permutation generator
1623 while ( $code->(@_[@idx]) ) {
1625 --$p while $idx[$p-1] > $idx[$p];
1626 my $q = $p or return;
1627 push @idx, reverse splice @idx, $p;
1628 ++$q while $idx[$p-1] > $idx[$q];
1629 @idx[$p-1,$q]=@idx[$q,$p-1];
1633 permute {print"@_\n"} split;
1635 =head2 How do I sort an array by (anything)?
1637 Supply a comparison function to sort() (described in L<perlfunc/sort>):
1639 @list = sort { $a <=> $b } @list;
1641 The default sort function is cmp, string comparison, which would
1642 sort C<(1, 2, 10)> into C<(1, 10, 2)>. C<< <=> >>, used above, is
1643 the numerical comparison operator.
1645 If you have a complicated function needed to pull out the part you
1646 want to sort on, then don't do it inside the sort function. Pull it
1647 out first, because the sort BLOCK can be called many times for the
1648 same element. Here's an example of how to pull out the first word
1649 after the first number on each item, and then sort those words
1654 ($item) = /\d+\s*(\S+)/;
1655 push @idx, uc($item);
1657 @sorted = @data[ sort { $idx[$a] cmp $idx[$b] } 0 .. $#idx ];
1659 which could also be written this way, using a trick
1660 that's come to be known as the Schwartzian Transform:
1662 @sorted = map { $_->[0] }
1663 sort { $a->[1] cmp $b->[1] }
1664 map { [ $_, uc( (/\d+\s*(\S+)/)[0]) ] } @data;
1666 If you need to sort on several fields, the following paradigm is useful.
1669 field1($a) <=> field1($b) ||
1670 field2($a) cmp field2($b) ||
1671 field3($a) cmp field3($b)
1674 This can be conveniently combined with precalculation of keys as given
1677 See the F<sort> article in the "Far More Than You Ever Wanted
1678 To Know" collection in http://www.cpan.org/misc/olddoc/FMTEYEWTK.tgz for
1679 more about this approach.
1681 See also the question later in L<perlfaq4> on sorting hashes.
1683 =head2 How do I manipulate arrays of bits?
1685 Use C<pack()> and C<unpack()>, or else C<vec()> and the bitwise
1688 For example, this sets C<$vec> to have bit N set if C<$ints[N]> was
1692 foreach(@ints) { vec($vec,$_,1) = 1 }
1694 Here's how, given a vector in C<$vec>, you can get those bits into your
1697 sub bitvec_to_list {
1700 # Find null-byte density then select best algorithm
1701 if ($vec =~ tr/\0// / length $vec > 0.95) {
1705 # This method is faster with mostly null-bytes
1706 while($vec =~ /[^\0]/g ) {
1707 $i = -9 + 8 * pos $vec;
1708 push @ints, $i if vec($vec, ++$i, 1);
1709 push @ints, $i if vec($vec, ++$i, 1);
1710 push @ints, $i if vec($vec, ++$i, 1);
1711 push @ints, $i if vec($vec, ++$i, 1);
1712 push @ints, $i if vec($vec, ++$i, 1);
1713 push @ints, $i if vec($vec, ++$i, 1);
1714 push @ints, $i if vec($vec, ++$i, 1);
1715 push @ints, $i if vec($vec, ++$i, 1);
1719 # This method is a fast general algorithm
1721 my $bits = unpack "b*", $vec;
1722 push @ints, 0 if $bits =~ s/^(\d)// && $1;
1723 push @ints, pos $bits while($bits =~ /1/g);
1729 This method gets faster the more sparse the bit vector is.
1730 (Courtesy of Tim Bunce and Winfried Koenig.)
1732 You can make the while loop a lot shorter with this suggestion
1733 from Benjamin Goldberg:
1735 while($vec =~ /[^\0]+/g ) {
1736 push @ints, grep vec($vec, $_, 1), $-[0] * 8 .. $+[0] * 8;
1739 Or use the CPAN module C<Bit::Vector>:
1741 $vector = Bit::Vector->new($num_of_bits);
1742 $vector->Index_List_Store(@ints);
1743 @ints = $vector->Index_List_Read();
1745 C<Bit::Vector> provides efficient methods for bit vector, sets of
1746 small integers and "big int" math.
1748 Here's a more extensive illustration using vec():
1751 $vector = "\xff\x0f\xef\xfe";
1752 print "Ilya's string \\xff\\x0f\\xef\\xfe represents the number ",
1753 unpack("N", $vector), "\n";
1754 $is_set = vec($vector, 23, 1);
1755 print "Its 23rd bit is ", $is_set ? "set" : "clear", ".\n";
1773 my ($offset, $width, $value) = @_;
1775 vec($vector, $offset, $width) = $value;
1776 print "offset=$offset width=$width value=$value\n";
1782 my $bits = unpack("b*", $vector);
1786 print "vector length in bytes: ", length($vector), "\n";
1787 @bytes = unpack("A8" x length($vector), $bits);
1788 print "bits are: @bytes\n\n";
1791 =head2 Why does defined() return true on empty arrays and hashes?
1793 The short story is that you should probably only use defined on scalars or
1794 functions, not on aggregates (arrays and hashes). See L<perlfunc/defined>
1795 in the 5.004 release or later of Perl for more detail.
1797 =head1 Data: Hashes (Associative Arrays)
1799 =head2 How do I process an entire hash?
1801 (contributed by brian d foy)
1803 There are a couple of ways that you can process an entire hash. You
1804 can get a list of keys, then go through each key, or grab a one
1805 key-value pair at a time.
1807 To go through all of the keys, use the C<keys> function. This extracts
1808 all of the keys of the hash and gives them back to you as a list. You
1809 can then get the value through the particular key you're processing:
1811 foreach my $key ( keys %hash ) {
1812 my $value = $hash{$key}
1816 Once you have the list of keys, you can process that list before you
1817 process the hashh elements. For instance, you can sort the keys so you
1818 can process them in lexical order:
1820 foreach my $key ( sort keys %hash ) {
1821 my $value = $hash{$key}
1825 Or, you might want to only process some of the items. If you only want
1826 to deal with the keys that start with C<text:>, you can select just
1827 those using C<grep>:
1829 foreach my $key ( grep /^text:/, keys %hash ) {
1830 my $value = $hash{$key}
1834 If the hash is very large, you might not want to create a long list of
1835 keys. To save some memory, you can grab on key-value pair at a time using
1836 C<each()>, which returns a pair you haven't seen yet:
1838 while( my( $key, $value ) = each( %hash ) ) {
1842 The C<each> operator returns the pairs in apparently random order, so if
1843 ordering matters to you, you'll have to stick with the C<keys> method.
1845 The C<each()> operator can be a bit tricky though. You can't add or
1846 delete keys of the hash while you're using it without possibly
1847 skipping or re-processing some pairs after Perl internally rehashes
1848 all of the elements. Additionally, a hash has only one iterator, so if
1849 you use C<keys>, C<values>, or C<each> on the same hash, you can reset
1850 the iterator and mess up your processing. See the C<each> entry in
1851 L<perlfunc> for more details.
1853 =head2 What happens if I add or remove keys from a hash while iterating over it?
1855 (contributed by brian d foy)
1857 The easy answer is "Don't do that!"
1859 If you iterate through the hash with each(), you can delete the key
1860 most recently returned without worrying about it. If you delete or add
1861 other keys, the iterator may skip or double up on them since perl
1862 may rearrange the hash table. See the
1863 entry for C<each()> in L<perlfunc>.
1865 =head2 How do I look up a hash element by value?
1867 Create a reverse hash:
1869 %by_value = reverse %by_key;
1870 $key = $by_value{$value};
1872 That's not particularly efficient. It would be more space-efficient
1875 while (($key, $value) = each %by_key) {
1876 $by_value{$value} = $key;
1879 If your hash could have repeated values, the methods above will only find
1880 one of the associated keys. This may or may not worry you. If it does
1881 worry you, you can always reverse the hash into a hash of arrays instead:
1883 while (($key, $value) = each %by_key) {
1884 push @{$key_list_by_value{$value}}, $key;
1887 =head2 How can I know how many entries are in a hash?
1889 If you mean how many keys, then all you have to do is
1890 use the keys() function in a scalar context:
1892 $num_keys = keys %hash;
1894 The keys() function also resets the iterator, which means that you may
1895 see strange results if you use this between uses of other hash operators
1898 =head2 How do I sort a hash (optionally by value instead of key)?
1900 (contributed by brian d foy)
1902 To sort a hash, start with the keys. In this example, we give the list of
1903 keys to the sort function which then compares them ASCIIbetically (which
1904 might be affected by your locale settings). The output list has the keys
1905 in ASCIIbetical order. Once we have the keys, we can go through them to
1906 create a report which lists the keys in ASCIIbetical order.
1908 my @keys = sort { $a cmp $b } keys %hash;
1910 foreach my $key ( @keys )
1912 printf "%-20s %6d\n", $key, $hash{$value};
1915 We could get more fancy in the C<sort()> block though. Instead of
1916 comparing the keys, we can compute a value with them and use that
1917 value as the comparison.
1919 For instance, to make our report order case-insensitive, we use
1920 the C<\L> sequence in a double-quoted string to make everything
1921 lowercase. The C<sort()> block then compares the lowercased
1922 values to determine in which order to put the keys.
1924 my @keys = sort { "\L$a" cmp "\L$b" } keys %hash;
1926 Note: if the computation is expensive or the hash has many elements,
1927 you may want to look at the Schwartzian Transform to cache the
1928 computation results.
1930 If we want to sort by the hash value instead, we use the hash key
1931 to look it up. We still get out a list of keys, but this time they
1932 are ordered by their value.
1934 my @keys = sort { $hash{$a} <=> $hash{$b} } keys %hash;
1936 From there we can get more complex. If the hash values are the same,
1937 we can provide a secondary sort on the hash key.
1940 $hash{$a} <=> $hash{$b}
1945 =head2 How can I always keep my hash sorted?
1946 X<hash tie sort DB_File Tie::IxHash>
1948 You can look into using the C<DB_File> module and C<tie()> using the
1949 C<$DB_BTREE> hash bindings as documented in L<DB_File/"In Memory
1950 Databases">. The C<Tie::IxHash> module from CPAN might also be
1951 instructive. Although this does keep your hash sorted, you might not
1952 like the slow down you suffer from the tie interface. Are you sure you
1955 =head2 What's the difference between "delete" and "undef" with hashes?
1957 Hashes contain pairs of scalars: the first is the key, the
1958 second is the value. The key will be coerced to a string,
1959 although the value can be any kind of scalar: string,
1960 number, or reference. If a key C<$key> is present in
1961 %hash, C<exists($hash{$key})> will return true. The value
1962 for a given key can be C<undef>, in which case
1963 C<$hash{$key}> will be C<undef> while C<exists $hash{$key}>
1964 will return true. This corresponds to (C<$key>, C<undef>)
1967 Pictures help... here's the C<%hash> table:
1977 And these conditions hold
1981 defined $hash{'d'} is true
1982 defined $hash{'a'} is true
1983 exists $hash{'a'} is true (Perl5 only)
1984 grep ($_ eq 'a', keys %hash) is true
1990 your table now reads:
2001 and these conditions now hold; changes in caps:
2005 defined $hash{'d'} is true
2006 defined $hash{'a'} is FALSE
2007 exists $hash{'a'} is true (Perl5 only)
2008 grep ($_ eq 'a', keys %hash) is true
2010 Notice the last two: you have an undef value, but a defined key!
2016 your table now reads:
2025 and these conditions now hold; changes in caps:
2029 defined $hash{'d'} is true
2030 defined $hash{'a'} is false
2031 exists $hash{'a'} is FALSE (Perl5 only)
2032 grep ($_ eq 'a', keys %hash) is FALSE
2034 See, the whole entry is gone!
2036 =head2 Why don't my tied hashes make the defined/exists distinction?
2038 This depends on the tied hash's implementation of EXISTS().
2039 For example, there isn't the concept of undef with hashes
2040 that are tied to DBM* files. It also means that exists() and
2041 defined() do the same thing with a DBM* file, and what they
2042 end up doing is not what they do with ordinary hashes.
2044 =head2 How do I reset an each() operation part-way through?
2046 Using C<keys %hash> in scalar context returns the number of keys in
2047 the hash I<and> resets the iterator associated with the hash. You may
2048 need to do this if you use C<last> to exit a loop early so that when
2049 you re-enter it, the hash iterator has been reset.
2051 =head2 How can I get the unique keys from two hashes?
2053 First you extract the keys from the hashes into lists, then solve
2054 the "removing duplicates" problem described above. For example:
2057 for $element (keys(%foo), keys(%bar)) {
2064 @uniq = keys %{{%foo,%bar}};
2066 Or if you really want to save space:
2069 while (defined ($key = each %foo)) {
2072 while (defined ($key = each %bar)) {
2077 =head2 How can I store a multidimensional array in a DBM file?
2079 Either stringify the structure yourself (no fun), or else
2080 get the MLDBM (which uses Data::Dumper) module from CPAN and layer
2081 it on top of either DB_File or GDBM_File.
2083 =head2 How can I make my hash remember the order I put elements into it?
2085 Use the C<Tie::IxHash> from CPAN.
2089 tie my %myhash, 'Tie::IxHash';
2091 for (my $i=0; $i<20; $i++) {
2095 my @keys = keys %myhash;
2096 # @keys = (0,1,2,3,...)
2098 =head2 Why does passing a subroutine an undefined element in a hash create it?
2100 If you say something like:
2102 somefunc($hash{"nonesuch key here"});
2104 Then that element "autovivifies"; that is, it springs into existence
2105 whether you store something there or not. That's because functions
2106 get scalars passed in by reference. If somefunc() modifies C<$_[0]>,
2107 it has to be ready to write it back into the caller's version.
2109 This has been fixed as of Perl5.004.
2111 Normally, merely accessing a key's value for a nonexistent key does
2112 I<not> cause that key to be forever there. This is different than
2115 =head2 How can I make the Perl equivalent of a C structure/C++ class/hash or array of hashes or arrays?
2117 Usually a hash ref, perhaps like this:
2122 TITLE => "deputy peon",
2125 PALS => [ "Norbert", "Rhys", "Phineas"],
2128 References are documented in L<perlref> and the upcoming L<perlreftut>.
2129 Examples of complex data structures are given in L<perldsc> and
2130 L<perllol>. Examples of structures and object-oriented classes are
2133 =head2 How can I use a reference as a hash key?
2135 (contributed by brian d foy)
2137 Hash keys are strings, so you can't really use a reference as the key.
2138 When you try to do that, perl turns the reference into its stringified
2139 form (for instance, C<HASH(0xDEADBEEF)>). From there you can't get
2140 back the reference from the stringified form, at least without doing
2141 some extra work on your own. Also remember that hash keys must be
2142 unique, but two different variables can store the same reference (and
2143 those variables can change later).
2145 The C<Tie::RefHash> module, which is distributed with perl, might be
2146 what you want. It handles that extra work.
2150 =head2 How do I handle binary data correctly?
2152 Perl is binary clean, so it can handle binary data just fine.
2153 On Windows or DOS, however, you have to use C<binmode> for binary
2154 files to avoid conversions for line endings. In general, you should
2155 use C<binmode> any time you want to work with binary data.
2157 Also see L<perlfunc/"binmode"> or L<perlopentut>.
2159 If you're concerned about 8-bit textual data then see L<perllocale>.
2160 If you want to deal with multibyte characters, however, there are
2161 some gotchas. See the section on Regular Expressions.
2163 =head2 How do I determine whether a scalar is a number/whole/integer/float?
2165 Assuming that you don't care about IEEE notations like "NaN" or
2166 "Infinity", you probably just want to use a regular expression.
2168 if (/\D/) { print "has nondigits\n" }
2169 if (/^\d+$/) { print "is a whole number\n" }
2170 if (/^-?\d+$/) { print "is an integer\n" }
2171 if (/^[+-]?\d+$/) { print "is a +/- integer\n" }
2172 if (/^-?\d+\.?\d*$/) { print "is a real number\n" }
2173 if (/^-?(?:\d+(?:\.\d*)?|\.\d+)$/) { print "is a decimal number\n" }
2174 if (/^([+-]?)(?=\d|\.\d)\d*(\.\d*)?([Ee]([+-]?\d+))?$/)
2175 { print "a C float\n" }
2177 There are also some commonly used modules for the task.
2178 L<Scalar::Util> (distributed with 5.8) provides access to perl's
2179 internal function C<looks_like_number> for determining whether a
2180 variable looks like a number. L<Data::Types> exports functions that
2181 validate data types using both the above and other regular
2182 expressions. Thirdly, there is C<Regexp::Common> which has regular
2183 expressions to match various types of numbers. Those three modules are
2184 available from the CPAN.
2186 If you're on a POSIX system, Perl supports the C<POSIX::strtod>
2187 function. Its semantics are somewhat cumbersome, so here's a
2188 C<getnum> wrapper function for more convenient access. This function
2189 takes a string and returns the number it found, or C<undef> for input
2190 that isn't a C float. The C<is_numeric> function is a front end to
2191 C<getnum> if you just want to say, "Is this a float?"
2194 use POSIX qw(strtod);
2199 my($num, $unparsed) = strtod($str);
2200 if (($str eq '') || ($unparsed != 0) || $!) {
2208 sub is_numeric { defined getnum($_[0]) }
2210 Or you could check out the L<String::Scanf> module on the CPAN
2211 instead. The C<POSIX> module (part of the standard Perl distribution)
2212 provides the C<strtod> and C<strtol> for converting strings to double
2213 and longs, respectively.
2215 =head2 How do I keep persistent data across program calls?
2217 For some specific applications, you can use one of the DBM modules.
2218 See L<AnyDBM_File>. More generically, you should consult the C<FreezeThaw>
2219 or C<Storable> modules from CPAN. Starting from Perl 5.8 C<Storable> is part
2220 of the standard distribution. Here's one example using C<Storable>'s C<store>
2221 and C<retrieve> functions:
2224 store(\%hash, "filename");
2227 $href = retrieve("filename"); # by ref
2228 %hash = %{ retrieve("filename") }; # direct to hash
2230 =head2 How do I print out or copy a recursive data structure?
2232 The C<Data::Dumper> module on CPAN (or the 5.005 release of Perl) is great
2233 for printing out data structures. The C<Storable> module on CPAN (or the
2234 5.8 release of Perl), provides a function called C<dclone> that recursively
2235 copies its argument.
2237 use Storable qw(dclone);
2240 Where C<$r1> can be a reference to any kind of data structure you'd like.
2241 It will be deeply copied. Because C<dclone> takes and returns references,
2242 you'd have to add extra punctuation if you had a hash of arrays that
2245 %newhash = %{ dclone(\%oldhash) };
2247 =head2 How do I define methods for every class/object?
2249 Use the C<UNIVERSAL> class (see L<UNIVERSAL>).
2251 =head2 How do I verify a credit card checksum?
2253 Get the C<Business::CreditCard> module from CPAN.
2255 =head2 How do I pack arrays of doubles or floats for XS code?
2257 The kgbpack.c code in the C<PGPLOT> module on CPAN does just this.
2258 If you're doing a lot of float or double processing, consider using
2259 the C<PDL> module from CPAN instead--it makes number-crunching easy.
2263 Revision: $Revision: 8539 $
2265 Date: $Date: 2007-01-11 00:07:14 +0100 (jeu, 11 jan 2007) $
2267 See L<perlfaq> for source control details and availability.
2269 =head1 AUTHOR AND COPYRIGHT
2271 Copyright (c) 1997-2007 Tom Christiansen, Nathan Torkington, and
2272 other authors as noted. All rights reserved.
2274 This documentation is free; you can redistribute it and/or modify it
2275 under the same terms as Perl itself.
2277 Irrespective of its distribution, all code examples in this file
2278 are hereby placed into the public domain. You are permitted and
2279 encouraged to use this code in your own programs for fun
2280 or for profit as you see fit. A simple comment in the code giving
2281 credit would be courteous but is not required.