3 perlfaq5 - Files and Formats ($Revision: 10126 $)
7 This section deals with I/O and the "f" issues: filehandles, flushing,
10 =head2 How do I flush/unbuffer an output filehandle? Why must I do this?
11 X<flush> X<buffer> X<unbuffer> X<autoflush>
13 Perl does not support truly unbuffered output (except insofar as you
14 can C<syswrite(OUT, $char, 1)>), although it does support is "command
15 buffering", in which a physical write is performed after every output
18 The C standard I/O library (stdio) normally buffers characters sent to
19 devices so that there isn't a system call for each byte. In most stdio
20 implementations, the type of output buffering and the size of the
21 buffer varies according to the type of device. Perl's C<print()> and
22 C<write()> functions normally buffer output, while C<syswrite()>
23 bypasses buffering all together.
25 If you want your output to be sent immediately when you execute
26 C<print()> or C<write()> (for instance, for some network protocols),
27 you must set the handle's autoflush flag. This flag is the Perl
28 variable C<$|> and when it is set to a true value, Perl will flush the
29 handle's buffer after each C<print()> or C<write()>. Setting C<$|>
30 affects buffering only for the currently selected default filehandle.
31 You choose this handle with the one argument C<select()> call (see
32 L<perlvar/$E<verbar>> and L<perlfunc/select>).
34 Use C<select()> to choose the desired handle, then set its
35 per-filehandle variables.
37 $old_fh = select(OUTPUT_HANDLE);
41 Some modules offer object-oriented access to handles and their
42 variables, although they may be overkill if this is the only thing you
43 do with them. You can use C<IO::Handle>:
46 open my( $printer ), ">", "/dev/printer"); # but is this?
47 $printer->autoflush(1);
49 or C<IO::Socket> (which inherits from C<IO::Handle>):
51 use IO::Socket; # this one is kinda a pipe?
52 my $sock = IO::Socket::INET->new( 'www.example.com:80' );
56 You can also flush an C<IO::Handle> object without setting
57 C<autoflush>. Call the C<flush> method to flush the buffer yourself:
60 open my( $printer ), ">", "/dev/printer");
61 $printer->flush; # one time flush
64 =head2 How do I change, delete, or insert a line in a file, or append to the beginning of a file?
67 (contributed by brian d foy)
69 The basic idea of inserting, changing, or deleting a line from a text
70 file involves reading and printing the file to the point you want to
71 make the change, making the change, then reading and printing the rest
72 of the file. Perl doesn't provide random access to lines (especially
73 since the record input separator, C<$/>, is mutable), although modules
74 such as C<Tie::File> can fake it.
76 A Perl program to do these tasks takes the basic form of opening a
77 file, printing its lines, then closing the file:
79 open my $in, '<', $file or die "Can't read old file: $!";
80 open my $out, '>', "$file.new" or die "Can't write new file: $!";
89 Within that basic form, add the parts that you need to insert, change,
92 To prepend lines to the beginning, print those lines before you enter
93 the loop that prints the existing lines.
95 open my $in, '<', $file or die "Can't read old file: $!";
96 open my $out, '>', "$file.new" or die "Can't write new file: $!";
98 print "# Add this line to the top\n"; # <--- HERE'S THE MAGIC
107 To change existing lines, insert the code to modify the lines inside
108 the C<while> loop. In this case, the code finds all lowercased
109 versions of "perl" and uppercases them. The happens for every line, so
110 be sure that you're supposed to do that on every line!
112 open my $in, '<', $file or die "Can't read old file: $!";
113 open my $out, '>', "$file.new" or die "Can't write new file: $!";
115 print "# Add this line to the top\n";
125 To change only a particular line, the input line number, C<$.>, is
126 useful. First read and print the lines up to the one you want to
127 change. Next, read the single line you want to change, change it, and
128 print it. After that, read the rest of the lines and print those:
130 while( <$in> ) # print the lines before the change
133 last if $. == 4; # line number before change
137 $line =~ s/\b(perl)\b/Perl/g;
140 while( <$in> ) # print the rest of the lines
145 To skip lines, use the looping controls. The C<next> in this example
146 skips comment lines, and the C<last> stops all processing once it
147 encounters either C<__END__> or C<__DATA__>.
151 next if /^\s+#/; # skip comment lines
152 last if /^__(END|DATA)__$/; # stop at end of code marker
156 Do the same sort of thing to delete a particular line by using C<next>
157 to skip the lines you don't want to show up in the output. This
158 example skips every fifth line:
166 If, for some odd reason, you really want to see the whole file at once
167 rather than processing line by line, you can slurp it in (as long as
168 you can fit the whole thing in memory!):
170 open my $in, '<', $file or die "Can't read old file: $!"
171 open my $out, '>', "$file.new" or die "Can't write new file: $!";
173 my @lines = do { local $/; <$in> }; # slurp!
179 Modules such as C<File::Slurp> and C<Tie::File> can help with that
180 too. If you can, however, avoid reading the entire file at once. Perl
181 won't give that memory back to the operating system until the process
184 You can also use Perl one-liners to modify a file in-place. The
185 following changes all 'Fred' to 'Barney' in F<inFile.txt>, overwriting
186 the file with the new contents. With the C<-p> switch, Perl wraps a
187 C<while> loop around the code you specify with C<-e>, and C<-i> turns
188 on in-place editing. The current line is in C<$_>. With C<-p>, Perl
189 automatically prints the value of C<$_> at the end of the loop. See
190 L<perlrun> for more details.
192 perl -pi -e 's/Fred/Barney/' inFile.txt
194 To make a backup of C<inFile.txt>, give C<-i> a file extension to add:
196 perl -pi.bak -e 's/Fred/Barney/' inFile.txt
198 To change only the fifth line, you can add a test checking C<$.>, the
199 input line number, then only perform the operation when the test
202 perl -pi -e 's/Fred/Barney/ if $. == 5' inFile.txt
204 To add lines before a certain line, you can add a line (or lines!)
205 before Perl prints C<$_>:
207 perl -pi -e 'print "Put before third line\n" if $. == 3' inFile.txt
209 You can even add a line to the beginning of a file, since the current
210 line prints at the end of the loop:
212 perl -pi -e 'print "Put before first line\n" if $. == 1' inFile.txt
214 To insert a line after one already in the file, use the C<-n> switch.
215 It's just like C<-p> except that it doesn't print C<$_> at the end of
216 the loop, so you have to do that yourself. In this case, print C<$_>
217 first, then print the line that you want to add.
219 perl -ni -e 'print; print "Put after fifth line\n" if $. == 5' inFile.txt
221 To delete lines, only print the ones that you want.
223 perl -ni -e 'print unless /d/' inFile.txt
227 perl -pi -e 'next unless /d/' inFile.txt
229 =head2 How do I count the number of lines in a file?
230 X<file, counting lines> X<lines> X<line>
232 One fairly efficient way is to count newlines in the file. The
233 following program uses a feature of tr///, as documented in L<perlop>.
234 If your text file doesn't end with a newline, then it's not really a
235 proper text file, so this may report one fewer line than you expect.
238 open(FILE, $filename) or die "Can't open `$filename': $!";
239 while (sysread FILE, $buffer, 4096) {
240 $lines += ($buffer =~ tr/\n//);
244 This assumes no funny games with newline translations.
246 =head2 How can I use Perl's C<-i> option from within a program?
249 C<-i> sets the value of Perl's C<$^I> variable, which in turn affects
250 the behavior of C<< <> >>; see L<perlrun> for more details. By
251 modifying the appropriate variables directly, you can get the same
252 behavior within a larger program. For example:
256 local($^I, @ARGV) = ('.orig', glob("*.c"));
259 print "This line should appear at the top of each file\n";
261 s/\b(p)earl\b/${1}erl/i; # Correct typos, preserving case
263 close ARGV if eof; # Reset $.
266 # $^I and @ARGV return to their old values here
268 This block modifies all the C<.c> files in the current directory,
269 leaving a backup of the original data from each file in a new
272 =head2 How can I copy a file?
273 X<copy> X<file, copy>
275 (contributed by brian d foy)
277 Use the File::Copy module. It comes with Perl and can do a
278 true copy across file systems, and it does its magic in
283 copy( $original, $new_copy ) or die "Copy failed: $!";
285 If you can't use File::Copy, you'll have to do the work yourself:
286 open the original file, open the destination file, then print
287 to the destination file as you read the original.
289 =head2 How do I make a temporary file name?
292 If you don't need to know the name of the file, you can use C<open()>
293 with C<undef> in place of the file name. The C<open()> function
294 creates an anonymous temporary file.
296 open my $tmp, '+>', undef or die $!;
298 Otherwise, you can use the File::Temp module.
300 use File::Temp qw/ tempfile tempdir /;
302 $dir = tempdir( CLEANUP => 1 );
303 ($fh, $filename) = tempfile( DIR => $dir );
305 # or if you don't need to know the filename
307 $fh = tempfile( DIR => $dir );
309 The File::Temp has been a standard module since Perl 5.6.1. If you
310 don't have a modern enough Perl installed, use the C<new_tmpfile>
311 class method from the IO::File module to get a filehandle opened for
312 reading and writing. Use it if you don't need to know the file's name:
315 $fh = IO::File->new_tmpfile()
316 or die "Unable to make new temporary file: $!";
318 If you're committed to creating a temporary file by hand, use the
319 process ID and/or the current time-value. If you need to have many
320 temporary files in one process, use a counter:
324 my $temp_dir = -d '/tmp' ? '/tmp' : $ENV{TMPDIR} || $ENV{TEMP};
325 my $base_name = sprintf "%s/%d-%d-0000", $temp_dir, $$, time;
330 until( defined(fileno(FH)) || $count++ > 100 ) {
331 $base_name =~ s/-(\d+)$/"-" . (1 + $1)/e;
332 # O_EXCL is required for security reasons.
333 sysopen FH, $base_name, O_WRONLY|O_EXCL|O_CREAT;
336 if( defined fileno(FH) ) {
337 return (*FH, $base_name);
346 =head2 How can I manipulate fixed-record-length files?
347 X<fixed-length> X<file, fixed-length records>
349 The most efficient way is using L<pack()|perlfunc/"pack"> and
350 L<unpack()|perlfunc/"unpack">. This is faster than using
351 L<substr()|perlfunc/"substr"> when taking many, many strings. It is
352 slower for just a few.
354 Here is a sample chunk of code to break up and put back together again
355 some fixed-format input lines, in this case from the output of a normal,
359 # 15158 p5 T 0:00 perl /home/tchrist/scripts/now-what
360 my $PS_T = 'A6 A4 A7 A5 A*';
361 open my $ps, '-|', 'ps';
363 my @fields = qw( pid tt stat time command );
366 @process{@fields} = unpack($PS_T, $_);
367 for my $field ( @fields ) {
368 print "$field: <$process{$field}>\n";
370 print 'line=', pack($PS_T, @process{@fields} ), "\n";
373 We've used a hash slice in order to easily handle the fields of each row.
374 Storing the keys in an array means it's easy to operate on them as a
375 group or loop over them with for. It also avoids polluting the program
376 with global variables and using symbolic references.
378 =head2 How can I make a filehandle local to a subroutine? How do I pass filehandles between subroutines? How do I make an array of filehandles?
379 X<filehandle, local> X<filehandle, passing> X<filehandle, reference>
381 As of perl5.6, open() autovivifies file and directory handles
382 as references if you pass it an uninitialized scalar variable.
383 You can then pass these references just like any other scalar,
384 and use them in the place of named handles.
386 open my $fh, $file_name;
388 open local $fh, $file_name;
390 print $fh "Hello World!\n";
394 If you like, you can store these filehandles in an array or a hash.
395 If you access them directly, they aren't simple scalars and you
396 need to give C<print> a little help by placing the filehandle
397 reference in braces. Perl can only figure it out on its own when
398 the filehandle reference is a simple scalar.
400 my @fhs = ( $fh1, $fh2, $fh3 );
402 for( $i = 0; $i <= $#fhs; $i++ ) {
403 print {$fhs[$i]} "just another Perl answer, \n";
406 Before perl5.6, you had to deal with various typeglob idioms
407 which you may see in older code.
409 open FILE, "> $filename";
410 process_typeglob( *FILE );
411 process_reference( \*FILE );
413 sub process_typeglob { local *FH = shift; print FH "Typeglob!" }
414 sub process_reference { local $fh = shift; print $fh "Reference!" }
416 If you want to create many anonymous handles, you should
417 check out the Symbol or IO::Handle modules.
419 =head2 How can I use a filehandle indirectly?
420 X<filehandle, indirect>
422 An indirect filehandle is using something other than a symbol
423 in a place that a filehandle is expected. Here are ways
424 to get indirect filehandles:
426 $fh = SOME_FH; # bareword is strict-subs hostile
427 $fh = "SOME_FH"; # strict-refs hostile; same package only
428 $fh = *SOME_FH; # typeglob
429 $fh = \*SOME_FH; # ref to typeglob (bless-able)
430 $fh = *SOME_FH{IO}; # blessed IO::Handle from *SOME_FH typeglob
432 Or, you can use the C<new> method from one of the IO::* modules to
433 create an anonymous filehandle, store that in a scalar variable,
434 and use it as though it were a normal filehandle.
436 use IO::Handle; # 5.004 or higher
437 $fh = IO::Handle->new();
439 Then use any of those as you would a normal filehandle. Anywhere that
440 Perl is expecting a filehandle, an indirect filehandle may be used
441 instead. An indirect filehandle is just a scalar variable that contains
442 a filehandle. Functions like C<print>, C<open>, C<seek>, or
443 the C<< <FH> >> diamond operator will accept either a named filehandle
444 or a scalar variable containing one:
446 ($ifh, $ofh, $efh) = (*STDIN, *STDOUT, *STDERR);
447 print $ofh "Type it: ";
449 print $efh "What was that: $got";
451 If you're passing a filehandle to a function, you can write
452 the function in two ways:
456 print $fh "Sending to indirect filehandle\n";
459 Or it can localize a typeglob and use the filehandle directly:
463 print FH "Sending to localized filehandle\n";
466 Both styles work with either objects or typeglobs of real filehandles.
467 (They might also work with strings under some circumstances, but this
473 In the examples above, we assigned the filehandle to a scalar variable
474 before using it. That is because only simple scalar variables, not
475 expressions or subscripts of hashes or arrays, can be used with
476 built-ins like C<print>, C<printf>, or the diamond operator. Using
477 something other than a simple scalar variable as a filehandle is
478 illegal and won't even compile:
480 @fd = (*STDIN, *STDOUT, *STDERR);
481 print $fd[1] "Type it: "; # WRONG
482 $got = <$fd[0]> # WRONG
483 print $fd[2] "What was that: $got"; # WRONG
485 With C<print> and C<printf>, you get around this by using a block and
486 an expression where you would place the filehandle:
488 print { $fd[1] } "funny stuff\n";
489 printf { $fd[1] } "Pity the poor %x.\n", 3_735_928_559;
490 # Pity the poor deadbeef.
492 That block is a proper block like any other, so you can put more
493 complicated code there. This sends the message out to one of two places:
496 print { $ok ? $fd[1] : $fd[2] } "cat stat $ok\n";
497 print { $fd[ 1+ ($ok || 0) ] } "cat stat $ok\n";
499 This approach of treating C<print> and C<printf> like object methods
500 calls doesn't work for the diamond operator. That's because it's a
501 real operator, not just a function with a comma-less argument. Assuming
502 you've been storing typeglobs in your structure as we did above, you
503 can use the built-in function named C<readline> to read a record just
504 as C<< <> >> does. Given the initialization shown above for @fd, this
505 would work, but only because readline() requires a typeglob. It doesn't
506 work with objects or strings, which might be a bug we haven't fixed yet.
508 $got = readline($fd[0]);
510 Let it be noted that the flakiness of indirect filehandles is not
511 related to whether they're strings, typeglobs, objects, or anything else.
512 It's the syntax of the fundamental operators. Playing the object
513 game doesn't help you at all here.
515 =head2 How can I set up a footer format to be used with write()?
518 There's no builtin way to do this, but L<perlform> has a couple of
519 techniques to make it possible for the intrepid hacker.
521 =head2 How can I write() into a string?
522 X<write, into a string>
524 See L<perlform/"Accessing Formatting Internals"> for an C<swrite()> function.
526 =head2 How can I open a filehandle to a string?
527 X<string>, X<open>, X<IO::Scalar>, X<filehandle>
529 (contributed by Peter J. Holzer, hjp-usenet2@hjp.at)
531 Since Perl 5.8.0, you can pass a reference to a scalar instead of the
532 filename to create a file handle which you can used to read from or write to
535 open(my $fh, '>', \$string) or die "Could not open string for writing";
537 print $fh "bar\n"; # $string now contains "foo\nbar\n"
539 open(my $fh, '<', \$string) or die "Could not open string for reading";
540 my $x = <$fh>; # $x now contains "foo\n"
542 With older versions of Perl, the C<IO::String> module provides similar
545 =head2 How can I output my numbers with commas added?
548 (contributed by brian d foy and Benjamin Goldberg)
550 You can use L<Number::Format> to separate places in a number.
551 It handles locale information for those of you who want to insert
552 full stops instead (or anything else that they want to use,
555 This subroutine will add commas to your number:
559 1 while s/^([-+]?\d+)(\d{3})/$1,$2/;
563 This regex from Benjamin Goldberg will add commas to numbers:
565 s/(^[-+]?\d+?(?=(?>(?:\d{3})+)(?!\d))|\G\d{3}(?=\d))/$1,/g;
567 It is easier to see with comments:
570 ^[-+]? # beginning of number.
571 \d+? # first digits before first comma
572 (?= # followed by, (but not included in the match) :
573 (?>(?:\d{3})+) # some positive multiple of three digits.
574 (?!\d) # an *exact* multiple, not x * 3 + 1 or whatever.
577 \G\d{3} # after the last group, get three digits
578 (?=\d) # but they have to have more digits after them.
581 =head2 How can I translate tildes (~) in a filename?
582 X<tilde> X<tilde expansion>
584 Use the <> (glob()) operator, documented in L<perlfunc>. Older
585 versions of Perl require that you have a shell installed that groks
586 tildes. Recent perl versions have this feature built in. The
587 File::KGlob module (available from CPAN) gives more portable glob
590 Within Perl, you may use this directly:
593 ^ ~ # find a leading tilde
595 [^/] # a non-slash character
596 * # repeated 0 or more times (0 means me)
601 : ( $ENV{HOME} || $ENV{LOGDIR} )
604 =head2 How come when I open a file read-write it wipes it out?
605 X<clobber> X<read-write> X<clobbering> X<truncate> X<truncating>
607 Because you're using something like this, which truncates the file and
608 I<then> gives you read-write access:
610 open(FH, "+> /path/name"); # WRONG (almost always)
612 Whoops. You should instead use this, which will fail if the file
615 open(FH, "+< /path/name"); # open for update
617 Using ">" always clobbers or creates. Using "<" never does
618 either. The "+" doesn't change this.
620 Here are examples of many kinds of file opens. Those using sysopen()
625 To open file for reading:
627 open(FH, "< $path") || die $!;
628 sysopen(FH, $path, O_RDONLY) || die $!;
630 To open file for writing, create new file if needed or else truncate old file:
632 open(FH, "> $path") || die $!;
633 sysopen(FH, $path, O_WRONLY|O_TRUNC|O_CREAT) || die $!;
634 sysopen(FH, $path, O_WRONLY|O_TRUNC|O_CREAT, 0666) || die $!;
636 To open file for writing, create new file, file must not exist:
638 sysopen(FH, $path, O_WRONLY|O_EXCL|O_CREAT) || die $!;
639 sysopen(FH, $path, O_WRONLY|O_EXCL|O_CREAT, 0666) || die $!;
641 To open file for appending, create if necessary:
643 open(FH, ">> $path") || die $!;
644 sysopen(FH, $path, O_WRONLY|O_APPEND|O_CREAT) || die $!;
645 sysopen(FH, $path, O_WRONLY|O_APPEND|O_CREAT, 0666) || die $!;
647 To open file for appending, file must exist:
649 sysopen(FH, $path, O_WRONLY|O_APPEND) || die $!;
651 To open file for update, file must exist:
653 open(FH, "+< $path") || die $!;
654 sysopen(FH, $path, O_RDWR) || die $!;
656 To open file for update, create file if necessary:
658 sysopen(FH, $path, O_RDWR|O_CREAT) || die $!;
659 sysopen(FH, $path, O_RDWR|O_CREAT, 0666) || die $!;
661 To open file for update, file must not exist:
663 sysopen(FH, $path, O_RDWR|O_EXCL|O_CREAT) || die $!;
664 sysopen(FH, $path, O_RDWR|O_EXCL|O_CREAT, 0666) || die $!;
666 To open a file without blocking, creating if necessary:
668 sysopen(FH, "/foo/somefile", O_WRONLY|O_NDELAY|O_CREAT)
669 or die "can't open /foo/somefile: $!":
671 Be warned that neither creation nor deletion of files is guaranteed to
672 be an atomic operation over NFS. That is, two processes might both
673 successfully create or unlink the same file! Therefore O_EXCL
674 isn't as exclusive as you might wish.
676 See also the new L<perlopentut> if you have it (new for 5.6).
678 =head2 Why do I sometimes get an "Argument list too long" when I use E<lt>*E<gt>?
679 X<argument list too long>
681 The C<< <> >> operator performs a globbing operation (see above).
682 In Perl versions earlier than v5.6.0, the internal glob() operator forks
683 csh(1) to do the actual glob expansion, but
684 csh can't handle more than 127 items and so gives the error message
685 C<Argument list too long>. People who installed tcsh as csh won't
686 have this problem, but their users may be surprised by it.
688 To get around this, either upgrade to Perl v5.6.0 or later, do the glob
689 yourself with readdir() and patterns, or use a module like File::KGlob,
690 one that doesn't use the shell to do globbing.
692 =head2 Is there a leak/bug in glob()?
695 Due to the current implementation on some operating systems, when you
696 use the glob() function or its angle-bracket alias in a scalar
697 context, you may cause a memory leak and/or unpredictable behavior. It's
698 best therefore to use glob() only in list context.
700 =head2 How can I open a file with a leading ">" or trailing blanks?
701 X<filename, special characters>
703 (contributed by Brian McCauley)
705 The special two argument form of Perl's open() function ignores
706 trailing blanks in filenames and infers the mode from certain leading
707 characters (or a trailing "|"). In older versions of Perl this was the
708 only version of open() and so it is prevalent in old code and books.
710 Unless you have a particular reason to use the two argument form you
711 should use the three argument form of open() which does not treat any
712 characters in the filename as special.
714 open FILE, "<", " file "; # filename is " file "
715 open FILE, ">", ">file"; # filename is ">file"
717 =head2 How can I reliably rename a file?
718 X<rename> X<mv> X<move> X<file, rename> X<ren>
720 If your operating system supports a proper mv(1) utility or its
721 functional equivalent, this works:
723 rename($old, $new) or system("mv", $old, $new);
725 It may be more portable to use the File::Copy module instead.
726 You just copy to the new file to the new name (checking return
727 values), then delete the old one. This isn't really the same
728 semantically as a rename(), which preserves meta-information like
729 permissions, timestamps, inode info, etc.
731 Newer versions of File::Copy export a move() function.
733 =head2 How can I lock a file?
734 X<lock> X<file, lock> X<flock>
736 Perl's builtin flock() function (see L<perlfunc> for details) will call
737 flock(2) if that exists, fcntl(2) if it doesn't (on perl version 5.004 and
738 later), and lockf(3) if neither of the two previous system calls exists.
739 On some systems, it may even use a different form of native locking.
740 Here are some gotchas with Perl's flock():
746 Produces a fatal error if none of the three system calls (or their
747 close equivalent) exists.
751 lockf(3) does not provide shared locking, and requires that the
752 filehandle be open for writing (or appending, or read/writing).
756 Some versions of flock() can't lock files over a network (e.g. on NFS file
757 systems), so you'd need to force the use of fcntl(2) when you build Perl.
758 But even this is dubious at best. See the flock entry of L<perlfunc>
759 and the F<INSTALL> file in the source distribution for information on
760 building Perl to do this.
762 Two potentially non-obvious but traditional flock semantics are that
763 it waits indefinitely until the lock is granted, and that its locks are
764 I<merely advisory>. Such discretionary locks are more flexible, but
765 offer fewer guarantees. This means that files locked with flock() may
766 be modified by programs that do not also use flock(). Cars that stop
767 for red lights get on well with each other, but not with cars that don't
768 stop for red lights. See the perlport manpage, your port's specific
769 documentation, or your system-specific local manpages for details. It's
770 best to assume traditional behavior if you're writing portable programs.
771 (If you're not, you should as always feel perfectly free to write
772 for your own system's idiosyncrasies (sometimes called "features").
773 Slavish adherence to portability concerns shouldn't get in the way of
774 your getting your job done.)
776 For more information on file locking, see also
777 L<perlopentut/"File Locking"> if you have it (new for 5.6).
781 =head2 Why can't I just open(FH, "E<gt>file.lock")?
782 X<lock, lockfile race condition>
784 A common bit of code B<NOT TO USE> is this:
786 sleep(3) while -e "file.lock"; # PLEASE DO NOT USE
787 open(LCK, "> file.lock"); # THIS BROKEN CODE
789 This is a classic race condition: you take two steps to do something
790 which must be done in one. That's why computer hardware provides an
791 atomic test-and-set instruction. In theory, this "ought" to work:
793 sysopen(FH, "file.lock", O_WRONLY|O_EXCL|O_CREAT)
794 or die "can't open file.lock: $!";
796 except that lamentably, file creation (and deletion) is not atomic
797 over NFS, so this won't work (at least, not every time) over the net.
798 Various schemes involving link() have been suggested, but
799 these tend to involve busy-wait, which is also less than desirable.
801 =head2 I still don't get locking. I just want to increment the number in the file. How can I do this?
802 X<counter> X<file, counter>
804 Didn't anyone ever tell you web-page hit counters were useless?
805 They don't count number of hits, they're a waste of time, and they serve
806 only to stroke the writer's vanity. It's better to pick a random number;
807 they're more realistic.
809 Anyway, this is what you can do if you can't help yourself.
811 use Fcntl qw(:DEFAULT :flock);
812 sysopen(FH, "numfile", O_RDWR|O_CREAT) or die "can't open numfile: $!";
813 flock(FH, LOCK_EX) or die "can't flock numfile: $!";
815 seek(FH, 0, 0) or die "can't rewind numfile: $!";
816 truncate(FH, 0) or die "can't truncate numfile: $!";
817 (print FH $num+1, "\n") or die "can't write numfile: $!";
818 close FH or die "can't close numfile: $!";
820 Here's a much better web-page hit counter:
822 $hits = int( (time() - 850_000_000) / rand(1_000) );
824 If the count doesn't impress your friends, then the code might. :-)
826 =head2 All I want to do is append a small amount of text to the end of a file. Do I still have to use locking?
827 X<append> X<file, append>
829 If you are on a system that correctly implements flock() and you use the
830 example appending code from "perldoc -f flock" everything will be OK
831 even if the OS you are on doesn't implement append mode correctly (if
832 such a system exists.) So if you are happy to restrict yourself to OSs
833 that implement flock() (and that's not really much of a restriction)
834 then that is what you should do.
836 If you know you are only going to use a system that does correctly
837 implement appending (i.e. not Win32) then you can omit the seek() from
838 the code in the previous answer.
840 If you know you are only writing code to run on an OS and filesystem that
841 does implement append mode correctly (a local filesystem on a modern
842 Unix for example), and you keep the file in block-buffered mode and you
843 write less than one buffer-full of output between each manual flushing
844 of the buffer then each bufferload is almost guaranteed to be written to
845 the end of the file in one chunk without getting intermingled with
846 anyone else's output. You can also use the syswrite() function which is
847 simply a wrapper around your systems write(2) system call.
849 There is still a small theoretical chance that a signal will interrupt
850 the system level write() operation before completion. There is also a
851 possibility that some STDIO implementations may call multiple system
852 level write()s even if the buffer was empty to start. There may be some
853 systems where this probability is reduced to zero.
855 =head2 How do I randomly update a binary file?
856 X<file, binary patch>
858 If you're just trying to patch a binary, in many cases something as
859 simple as this works:
861 perl -i -pe 's{window manager}{window mangler}g' /usr/bin/emacs
863 However, if you have fixed sized records, then you might do something more
866 $RECSIZE = 220; # size of record, in bytes
867 $recno = 37; # which record to update
868 open(FH, "+<somewhere") || die "can't update somewhere: $!";
869 seek(FH, $recno * $RECSIZE, 0);
870 read(FH, $record, $RECSIZE) == $RECSIZE || die "can't read record $recno: $!";
872 seek(FH, -$RECSIZE, 1);
876 Locking and error checking are left as an exercise for the reader.
877 Don't forget them or you'll be quite sorry.
879 =head2 How do I get a file's timestamp in perl?
880 X<timestamp> X<file, timestamp>
882 If you want to retrieve the time at which the file was last
883 read, written, or had its meta-data (owner, etc) changed,
884 you use the B<-A>, B<-M>, or B<-C> file test operations as
885 documented in L<perlfunc>. These retrieve the age of the
886 file (measured against the start-time of your program) in
887 days as a floating point number. Some platforms may not have
888 all of these times. See L<perlport> for details. To
889 retrieve the "raw" time in seconds since the epoch, you
890 would call the stat function, then use localtime(),
891 gmtime(), or POSIX::strftime() to convert this into
896 $write_secs = (stat($file))[9];
897 printf "file %s updated at %s\n", $file,
898 scalar localtime($write_secs);
900 If you prefer something more legible, use the File::stat module
901 (part of the standard distribution in version 5.004 and later):
903 # error checking left as an exercise for reader.
906 $date_string = ctime(stat($file)->mtime);
907 print "file $file updated at $date_string\n";
909 The POSIX::strftime() approach has the benefit of being,
910 in theory, independent of the current locale. See L<perllocale>
913 =head2 How do I set a file's timestamp in perl?
914 X<timestamp> X<file, timestamp>
916 You use the utime() function documented in L<perlfunc/utime>.
917 By way of example, here's a little program that copies the
918 read and write times from its first argument to all the rest
922 die "usage: cptimes timestamp_file other_files ...\n";
925 ($atime, $mtime) = (stat($timestamp))[8,9];
926 utime $atime, $mtime, @ARGV;
928 Error checking is, as usual, left as an exercise for the reader.
930 The perldoc for utime also has an example that has the same
931 effect as touch(1) on files that I<already exist>.
933 Certain file systems have a limited ability to store the times
934 on a file at the expected level of precision. For example, the
935 FAT and HPFS filesystem are unable to create dates on files with
936 a finer granularity than two seconds. This is a limitation of
937 the filesystems, not of utime().
939 =head2 How do I print to more than one file at once?
940 X<print, to multiple files>
942 To connect one filehandle to several output filehandles,
943 you can use the IO::Tee or Tie::FileHandle::Multiplex modules.
945 If you only have to do this once, you can print individually
948 for $fh (FH1, FH2, FH3) { print $fh "whatever\n" }
950 =head2 How can I read in an entire file all at once?
951 X<slurp> X<file, slurping>
953 You can use the File::Slurp module to do it in one step.
957 $all_of_it = read_file($filename); # entire file in scalar
958 @all_lines = read_file($filename); # one line perl element
960 The customary Perl approach for processing all the lines in a file is to
961 do so one line at a time:
963 open (INPUT, $file) || die "can't open $file: $!";
966 # do something with $_
968 close(INPUT) || die "can't close $file: $!";
970 This is tremendously more efficient than reading the entire file into
971 memory as an array of lines and then processing it one element at a time,
972 which is often--if not almost always--the wrong approach. Whenever
973 you see someone do this:
977 you should think long and hard about why you need everything loaded at
978 once. It's just not a scalable solution. You might also find it more
979 fun to use the standard Tie::File module, or the DB_File module's
980 $DB_RECNO bindings, which allow you to tie an array to a file so that
981 accessing an element the array actually accesses the corresponding
984 You can read the entire filehandle contents into a scalar.
988 open (INPUT, $file) || die "can't open $file: $!";
992 That temporarily undefs your record separator, and will automatically
993 close the file at block exit. If the file is already open, just use this:
995 $var = do { local $/; <INPUT> };
997 For ordinary files you can also use the read function.
999 read( INPUT, $var, -s INPUT );
1001 The third argument tests the byte size of the data on the INPUT filehandle
1002 and reads that many bytes into the buffer $var.
1004 =head2 How can I read in a file by paragraphs?
1005 X<file, reading by paragraphs>
1007 Use the C<$/> variable (see L<perlvar> for details). You can either
1008 set it to C<""> to eliminate empty paragraphs (C<"abc\n\n\n\ndef">,
1009 for instance, gets treated as two paragraphs and not three), or
1010 C<"\n\n"> to accept empty paragraphs.
1012 Note that a blank line must have no blanks in it. Thus
1013 S<C<"fred\n \nstuff\n\n">> is one paragraph, but C<"fred\n\nstuff\n\n"> is two.
1015 =head2 How can I read a single character from a file? From the keyboard?
1016 X<getc> X<file, reading one character at a time>
1018 You can use the builtin C<getc()> function for most filehandles, but
1019 it won't (easily) work on a terminal device. For STDIN, either use
1020 the Term::ReadKey module from CPAN or use the sample code in
1023 If your system supports the portable operating system programming
1024 interface (POSIX), you can use the following code, which you'll note
1025 turns off echo processing as well.
1039 use POSIX qw(:termios_h);
1041 my ($term, $oterm, $echo, $noecho, $fd_stdin);
1043 $fd_stdin = fileno(STDIN);
1045 $term = POSIX::Termios->new();
1046 $term->getattr($fd_stdin);
1047 $oterm = $term->getlflag();
1049 $echo = ECHO | ECHOK | ICANON;
1050 $noecho = $oterm & ~$echo;
1053 $term->setlflag($noecho);
1054 $term->setcc(VTIME, 1);
1055 $term->setattr($fd_stdin, TCSANOW);
1059 $term->setlflag($oterm);
1060 $term->setcc(VTIME, 0);
1061 $term->setattr($fd_stdin, TCSANOW);
1067 sysread(STDIN, $key, 1);
1076 The Term::ReadKey module from CPAN may be easier to use. Recent versions
1077 include also support for non-portable systems as well.
1080 open(TTY, "</dev/tty");
1081 print "Gimme a char: ";
1083 $key = ReadKey 0, *TTY;
1085 printf "\nYou said %s, char number %03d\n",
1088 =head2 How can I tell whether there's a character waiting on a filehandle?
1090 The very first thing you should do is look into getting the Term::ReadKey
1091 extension from CPAN. As we mentioned earlier, it now even has limited
1092 support for non-portable (read: not open systems, closed, proprietary,
1093 not POSIX, not Unix, etc) systems.
1095 You should also check out the Frequently Asked Questions list in
1096 comp.unix.* for things like this: the answer is essentially the same.
1097 It's very system dependent. Here's one solution that works on BSD
1102 vec($rin, fileno(STDIN), 1) = 1;
1103 return $nfd = select($rin,undef,undef,0);
1106 If you want to find out how many characters are waiting, there's
1107 also the FIONREAD ioctl call to be looked at. The I<h2ph> tool that
1108 comes with Perl tries to convert C include files to Perl code, which
1109 can be C<require>d. FIONREAD ends up defined as a function in the
1110 I<sys/ioctl.ph> file:
1112 require 'sys/ioctl.ph';
1114 $size = pack("L", 0);
1115 ioctl(FH, FIONREAD(), $size) or die "Couldn't call ioctl: $!\n";
1116 $size = unpack("L", $size);
1118 If I<h2ph> wasn't installed or doesn't work for you, you can
1119 I<grep> the include files by hand:
1121 % grep FIONREAD /usr/include/*/*
1122 /usr/include/asm/ioctls.h:#define FIONREAD 0x541B
1124 Or write a small C program using the editor of champions:
1127 #include <sys/ioctl.h>
1129 printf("%#08x\n", FIONREAD);
1132 % cc -o fionread fionread.c
1136 And then hard code it, leaving porting as an exercise to your successor.
1138 $FIONREAD = 0x4004667f; # XXX: opsys dependent
1140 $size = pack("L", 0);
1141 ioctl(FH, $FIONREAD, $size) or die "Couldn't call ioctl: $!\n";
1142 $size = unpack("L", $size);
1144 FIONREAD requires a filehandle connected to a stream, meaning that sockets,
1145 pipes, and tty devices work, but I<not> files.
1147 =head2 How do I do a C<tail -f> in perl?
1148 X<tail> X<IO::Handle> X<File::Tail> X<clearerr>
1154 The statement C<seek(GWFILE, 0, 1)> doesn't change the current position,
1155 but it does clear the end-of-file condition on the handle, so that the
1156 next C<< <GWFILE> >> makes Perl try again to read something.
1158 If that doesn't work (it relies on features of your stdio implementation),
1159 then you need something more like this:
1162 for ($curpos = tell(GWFILE); <GWFILE>; $curpos = tell(GWFILE)) {
1163 # search for some stuff and put it into files
1166 seek(GWFILE, $curpos, 0); # seek to where we had been
1169 If this still doesn't work, look into the C<clearerr> method
1170 from C<IO::Handle>, which resets the error and end-of-file states
1173 There's also a C<File::Tail> module from CPAN.
1175 =head2 How do I dup() a filehandle in Perl?
1178 If you check L<perlfunc/open>, you'll see that several of the ways
1179 to call open() should do the trick. For example:
1181 open(LOG, ">>/foo/logfile");
1182 open(STDERR, ">&LOG");
1184 Or even with a literal numeric descriptor:
1186 $fd = $ENV{MHCONTEXTFD};
1187 open(MHCONTEXT, "<&=$fd"); # like fdopen(3S)
1189 Note that "<&STDIN" makes a copy, but "<&=STDIN" make
1190 an alias. That means if you close an aliased handle, all
1191 aliases become inaccessible. This is not true with
1194 Error checking, as always, has been left as an exercise for the reader.
1196 =head2 How do I close a file descriptor by number?
1197 X<file, closing file descriptors> X<POSIX> X<close>
1199 If, for some reason, you have a file descriptor instead of a
1200 filehandle (perhaps you used C<POSIX::open>), you can use the
1201 C<close()> function from the C<POSIX> module:
1205 POSIX::close( $fd );
1207 This should rarely be necessary, as the Perl C<close()> function is to be
1208 used for things that Perl opened itself, even if it was a dup of a
1209 numeric descriptor as with C<MHCONTEXT> above. But if you really have
1210 to, you may be able to do this:
1212 require 'sys/syscall.ph';
1213 $rc = syscall(&SYS_close, $fd + 0); # must force numeric
1214 die "can't sysclose $fd: $!" unless $rc == -1;
1216 Or, just use the fdopen(3S) feature of C<open()>:
1219 open my( $fh ), "<&=$fd" or die "Cannot reopen fd=$fd: $!";
1223 =head2 Why can't I use "C:\temp\foo" in DOS paths? Why doesn't `C:\temp\foo.exe` work?
1224 X<filename, DOS issues>
1226 Whoops! You just put a tab and a formfeed into that filename!
1227 Remember that within double quoted strings ("like\this"), the
1228 backslash is an escape character. The full list of these is in
1229 L<perlop/Quote and Quote-like Operators>. Unsurprisingly, you don't
1230 have a file called "c:(tab)emp(formfeed)oo" or
1231 "c:(tab)emp(formfeed)oo.exe" on your legacy DOS filesystem.
1233 Either single-quote your strings, or (preferably) use forward slashes.
1234 Since all DOS and Windows versions since something like MS-DOS 2.0 or so
1235 have treated C</> and C<\> the same in a path, you might as well use the
1236 one that doesn't clash with Perl--or the POSIX shell, ANSI C and C++,
1237 awk, Tcl, Java, or Python, just to mention a few. POSIX paths
1238 are more portable, too.
1240 =head2 Why doesn't glob("*.*") get all the files?
1243 Because even on non-Unix ports, Perl's glob function follows standard
1244 Unix globbing semantics. You'll need C<glob("*")> to get all (non-hidden)
1245 files. This makes glob() portable even to legacy systems. Your
1246 port may include proprietary globbing functions as well. Check its
1247 documentation for details.
1249 =head2 Why does Perl let me delete read-only files? Why does C<-i> clobber protected files? Isn't this a bug in Perl?
1251 This is elaborately and painstakingly described in the
1252 F<file-dir-perms> article in the "Far More Than You Ever Wanted To
1253 Know" collection in http://www.cpan.org/misc/olddoc/FMTEYEWTK.tgz .
1255 The executive summary: learn how your filesystem works. The
1256 permissions on a file say what can happen to the data in that file.
1257 The permissions on a directory say what can happen to the list of
1258 files in that directory. If you delete a file, you're removing its
1259 name from the directory (so the operation depends on the permissions
1260 of the directory, not of the file). If you try to write to the file,
1261 the permissions of the file govern whether you're allowed to.
1263 =head2 How do I select a random line from a file?
1264 X<file, selecting a random line>
1266 Here's an algorithm from the Camel Book:
1269 rand($.) < 1 && ($line = $_) while <>;
1271 This has a significant advantage in space over reading the whole file
1272 in. You can find a proof of this method in I<The Art of Computer
1273 Programming>, Volume 2, Section 3.4.2, by Donald E. Knuth.
1275 You can use the File::Random module which provides a function
1278 use File::Random qw/random_line/;
1279 my $line = random_line($filename);
1281 Another way is to use the Tie::File module, which treats the entire
1282 file as an array. Simply access a random array element.
1284 =head2 Why do I get weird spaces when I print an array of lines?
1290 joins together the elements of C<@lines> with a space between them.
1291 If C<@lines> were C<("little", "fluffy", "clouds")> then the above
1292 statement would print
1294 little fluffy clouds
1296 but if each element of C<@lines> was a line of text, ending a newline
1297 character C<("little\n", "fluffy\n", "clouds\n")> then it would print:
1303 If your array contains lines, just print them:
1309 Revision: $Revision: 10126 $
1311 Date: $Date: 2007-10-27 21:29:20 +0200 (Sat, 27 Oct 2007) $
1313 See L<perlfaq> for source control details and availability.
1315 =head1 AUTHOR AND COPYRIGHT
1317 Copyright (c) 1997-2007 Tom Christiansen, Nathan Torkington, and
1318 other authors as noted. All rights reserved.
1320 This documentation is free; you can redistribute it and/or modify it
1321 under the same terms as Perl itself.
1323 Irrespective of its distribution, all code examples here are in the public
1324 domain. You are permitted and encouraged to use this code and any
1325 derivatives thereof in your own programs for fun or for profit as you
1326 see fit. A simple comment in the code giving credit to the FAQ would
1327 be courteous but is not required.