I'm working in a program where the user can pass a -o file option, and output should be then directed to that file. Otherwise, it should go to stdout.
To retrieve the option I'm using the module getopt long, and that's not the problem. The problem is that I want to create a file handle with that file or assign stdout to it if the option was not set.
if ($opt) {
open OUTPUT, ">", $file;
} else {
open OUTPUT, # ???
}
That's because this way, later in my code I can just:
print OUTPUT "...";
Without worrying if OUTPUT is stdout or a file the user specified. Is this possible? If I'm doing a bad design here, please let me know.
This would be a good example on how to use select.
use strict;
use warnings;
use autodie;
my $fh;
if ($opt) {
open $fh, '>', $file;
select $fh;
}
print "This goes to the file if $opt is defined, otherwise to STDOUT."
Look at the open documentation. The easiest is to reopen STDOUT itself and not use a filehandle in your code.
if ($opt) {
open(STDOUT, ">", $file);
}
...
print "this goes to $file or STDOUT\n";
(Add some error checking of course.)
A constant item such as OUTPUT cannot be assigned. Using a variable such as $output works better. For example:
my ($output, $display_filename);
if ($opt)
{
if ($opt eq '-')
{
$display_filename = 'stdout';
$output = *STDOUT;
}
else
{
$display_filename = $opt;
open($output, '>', $opt) or
die("Cannot open $opt for writing: $!\n");
}
}
That way the program can print to standard output and/or to an output file:
print $output "This might go to a file\n";
print "Data written to $display_filename\n" if ($verbose);
Related
I've looked around here a bit and found similar questions but not exactly. If there is one, I apologize and please point me to it.
I have the following code. I'm trying to create a csv file of simply an ID pulled from a filename and the filename itself. This is the ENTIRE script.
use strict;
use warnings;
use File::Find;
find( \&findAllFiles, '.');
exit;
sub findAllFiles {
my #fp1;
my #fp2;
my $patId;
my $filename;
my $testvar = "hello again";
$filename = $File::Find::name;
if ($filename =~ /\.pdf$/) {
open (my $fh, '>', 'filenames.csv') or die "Failed to open - $!\n";
print $fh "starting...$testvar\n" or die "Failed to print to file - $!\n";
#fp1 = split('/', $filename);
#fp2 = split('_', $fp1[-1]);
$patId = $fp2[-1];
$patId =~ s/\.pdf$//;
print "Adding $patId, file = $filename\n";
print $fh "$patId,$filename\n" or die "File print error: $!";
close $fh or warn "close failed! - $!";
}
return;
}
The line that prints to the screen, prints perfectly.
If I take the file open/close and the first print statement out of the if block, it prints that line into the file, but not the data inside the block.
I've tried every combo I can think of and it doesn't work. I've alternated between '>' and '>>' since it clearly needs the append since it's looping over filenames, but neither works inside the if block.
Even this code above doesn't throw the die errors! It just ignores those lines! I'm figuring there's something obvious I'm missing.
Quoting File::Find::find's documentation:
Additionally, for each directory found, it will chdir() into that directory
It means that when you open inside findAllFiles, you are potentially opening a file filenames.csv inside a subdirectory of your initial directory. You can run something like find . -name filenames.csv from your terminal, and you'll see plenty of filenames.csv. You can change this behavior by passing no_chdir option to find:
find( { wanted => \&findAllFiles, no_chdir => 1}, '.');
(and additionally changing > for >> in your open)
However, personally, I'd avoid repeatedly opening and closing filenames.csv when you could open it just once before calling find. If you don't want to have your filehandle globally defined, you can always pass it as an argument to findAllFiles:
{
open my $fh, '>', 'filenames.csv' or die "Failed to open 'filenames.csv': $!";
find(sub { findAllFiles($fh) }, '.')
}
sub findAllFiles {
my ($fh) = #_;
...
filenames.csv will be created in the directory where the pdf is found, since find() changes directories as it searches. If that's not what you want, use an absolute path to open it (or open it before calling find, which seems like a better idea).
I am trying this in Perl to pick one complete line from whole document which contains "CURRENT_RUN_ID". I have been using below code to accomplish the above said task but I am unable to enter the while loop.
my $sSuccessString = "CURRENT_RUN_ID";
open(LOG, "$slogfile") or die("Can't open $slogfile\n");
my $sLines;
{
local $/ = undef;
$sLines=<LOG>;
}
my $spool = 0;
my #matchingLines;
while (<LOG>)
{
print OUTLOG "in while loop\n";
if (m/$sSuccessString/i) {
print OUTLOG "in if loop\n";
$spool = 1;
print map { "$_ \n" } #matchingLines;
#matchingLines = ();
}
if ($spool) {
push (#matchingLines, $_);
}
}
You are already done reading from the filehandle LOG after you have slurped it into $sLines. <LOG> in the head of the while will return undef because it has reached eof. You either have to use that variable $sLines in your while loop or get rid of it. You're not using it anyway.
If you only want to print the line that matches, all you need to do is this:
use strict;
use warnings;
open my $fh_in, '<', 'input_file' or die $!;
open my $fh_out '>', 'output_file' or die $!;
while (my $line = <$fh_in>) {
print $fh_out $line if $line =~ m/CURRENT_RUN_ID/;
}
close $fh_in;
close $fh_out;
When you execute this code:
$sLines=<LOG>;
it reads all of the data from LOG into $sLines and it leaves the file pointer for LOG at the end of the file. So when you next try to read from that file handle with:
while (<LOG>)
nothing is returned as there is no more data to read.
If you want to read the file twice, then you will need to use the seek() function to reset the file pointer before your second read.
seek LOG, 0, 0;
But, given that you never do anything with $sLines I suspect that you can probably just remove that whole section of the code.
The whole thing with $spool and #matchingLines seems strange too. What were you trying to achieve there?
I think your code can be simplified to just:
my $sSuccessString = "CURRENT_RUN_ID";
open(LOG, $slogfile) or die("Can't open $slogfile\n");
while (<LOG>) {
print OUTLOG if /$sSuccessString/i/;
}
Personally, I'd make it even simpler, by reading from STDIN and writing to STDOUT.
my $sSuccessString = 'CURRENT_RUN_ID';
while (<>) {
print if /$sSuccessString/i/;
}
And then using Unix I/O redirection to connect up the correct files.
$ ./this_filter.pl < your_input.log > your_output.log
I want to make a tool to classify each line in input file to several files
but it seems have some problem in naming a filehandler so I can't go ahead , how do I solve?
here is my program
ARGV[0] is the input file
ARGV[1] is the number of classes
#!/usr/bin/perl
use POSIX;
use warnings;
# open input file
open(Raw,"<","./$ARGV[0]") or die "Can't open $ARGV[0] \n";
# create a directory class to store class files
system("mkdir","Class");
# create files for store class informations
for($i=1;$i<=$ARGV[1];$i++)
{
# it seems something wrong in here
open("Class$i",">","./Class/$i.class") or die "Can't create $i.class \n";
}
# read each line and random decide which class to store
while( eof(Raw) != 1)
{
$Line = readline(*Raw);
$Random_num = ceil(rand $ARGV[1]);
for($k=1;$k<=$ARGV[1];$k++)
{
if($Random_num == $k)
{
# Store to the file
print "Class$k" $Line;
last;
}
}
}
for($h=1;$h<=$ARGV[1];$h++)
{
close "Class$h";
}
close Raw;
thanks
Later I use the advice provided by Bill Ruppert
I put the name of filehandler into array , but it seems appear a syntax bug , but I can't correct it
I label the syntax bug with ######## A syntax error but it looks quite OK ########
here is my code
#!/usr/bin/perl
use POSIX;
use warnings;
use Data::Dumper;
# open input file
open(Raw,"<","./$ARGV[0]") or die "Can't open $ARGV[0] \n";
# create a directory class to store class files
system("mkdir","Class");
# put the name of hilehandler into array
for($i=0;$i<$ARGV[1];$i++)
{
push(#Name,("Class".$i));
}
# create files of classes
for($i=0;$i<=$#Name;$i++)
{
$I = ($i+1);
open($Name[$i],">","./Class/$I.class") or die "Can't create $I.class \n";
}
# read each line and random decide which class to store
while( eof(Raw) != 1)
{
$Line = readline(*Raw);
$Random_num = ceil(rand $ARGV[1]);
for($k=0;$k<=$#Name;$k++)
{
if($Random_num == ($k+1))
{
print $Name[$k] $Line; ######## A syntax error but it looks quite OK ########
last;
}
}
}
for($h=0;$h<=$#Name;$h++)
{
close $Name[$h];
}
close Raw;
thanks
To quote the Perl documentation on the print function:
If you're storing handles in an array or hash, or in general whenever you're using any expression more complex than a bareword handle or a plain, unsubscripted scalar variable to retrieve it, you will have to use a block returning the filehandle value instead, in which case the LIST may not be omitted:
print { $files[$i] } "stuff\n";
print { $OK ? STDOUT : STDERR } "stuff\n";
Thus, print $Name[$k] $Line; needs to be changed to print { $Name[$k] } $Line;.
How about this one:
#! /usr/bin/perl -w
use strict;
use POSIX;
my $input_file = shift;
my $file_count = shift;
my %hash;
open(INPUT, "<$input_file") || die "Can't open file $input_file";
while(my $line = <INPUT>) {
my $num = ceil(rand($file_count));
$hash{$num} .= $line
}
foreach my $i (1..$file_count) {
open(OUTPUT, ">$i.txt") || die "Can't open file $i.txt";
print OUTPUT $hash{$i};
close OUTPUT;
}
close INPUT;
I want to redirect STDERR and STDOUT to a variable. I did this.
close(STDOUT);
close(STDERR);
my $out;
open(STDOUT, ">>", \$out);
open(STDERR, ">>", \$out);
for(1..10)
{
print "print\n"; # this is ok.
warn "warn\n"; # same
system("make"); # this is lost. neither in screen nor in variable.
}
The problem with system. I want the output of this call to be captured too.
use Capture::Tiny!
Are you seeking to capture the output in a variable? If so, you have use backticks or qx{} with appropriate redirection. For example, you could use:
#/usr/bin/env perl
use strict;
use warnings;
# Ensure we have a way to write messages
open my $fh, '>', "output" or die;
close(STDOUT);
close(STDERR);
my $out;
open(STDOUT, ">>", \$out) or do { print $fh, "failed to open STDOUT ($!)\n"; die };
open(STDERR, ">>", \$out) or do { print $fh, "failed to open STDERR ($!)\n"; die };
foreach my $i (1..10)
{
print "print $i\n";
warn "warn $i\n";
my $extra = qx{make pth$i 2>&1};
print $fh "<<$i>><<$out>><<$extra>>\n";
}
(I happen to have programs pth1, pth2 and pth3 in the directory - they were made OK; pth4 and above write errors to stderr; the redirection was necessary.)
You should always check the success of operations such as open().
Why is this necessary? Because writing to a variable requires the cooperation of the process doing the writing - and make doesn't know how to cooperate.
There are several ways to redirect and restore STDOUT. Some of them work with STDERR too. Here are my two favorites:
Using select:
my $out;
open my $fh, ">>", \$out;
select $fh;
print "written to the variable\n";
select STDOUT;
print "written to original STDOUT\n";
Using local:
my $out
do {
local *STDOUT;
open STDOUT, ">>", \$out;
print "written to the variable\n";
};
print "written to original STDOUT\n";
Enjoy.
The reason this is happening is that the STDOUT and STDERR "filehandles" are not equivalent to stderr and stdout handles provided by the shell to the perl binary. In order to achieve what you want, you should use open instead of system
Why not use IPC::Open3?
TLDR Answer
use Capture::Tiny;
Merged STDOUT and STDERR
If you want STDOUT (from print()s) and STDERR (from warn()s) to be merged, then use...
my ($merged, #result) = capture_merged { print "Hello, world!" }; # static code
my ($merged, #result) = capture_merged { eval $codetoeval }; # code in variable
Separated STDOUT and STDERR
If you want them separated...
my ($stdout, $stderr, #result) = capture { print "Hello, world!" }; # static code
my ($stdout, $stderr, #result) = capture { eval $codetoeval }; # code in variable
Results of Eval
#result indicates the success, with success being [1], and failure being []. Tiny has a ton of other functions that you can look through for other cases, like code references, etc.. But I think the code above should cover most of any Perl developer's needs.
For some reason, I can't get filehandles working with Expect.pm's log_file method. I originally got help on How can I pass a filehandle to Perl Expect's log_file function?, where it was suggested that I use an IO::Handle filehandle to pass to the method. This seems to be a different issue, so I thought I'd start a new question.
This is the offending section of Expect.pm:
if (ref($file) ne 'CODE') {
croak "Given logfile doesn't have a 'print' method"
if not $fh->can("print");
$fh->autoflush(1); # so logfile is up to date
}
So, then, I tried this sample code:
use IO::Handle;
open $fh, ">>", "file.out" or die "Can't open file";
$fh->print("Hello, world");
if ($fh->can("print"))
{
print "Yes\n";
}
else
{
print "No\n";
}
When I run this, I get two (to my mind) conflicting items. A file with a single line that says 'Hello, world', and output of 'No'. To my mind, the $fh->can line should return true. Am I wrong here?
Odd, it looks like you need to create a real IO::File object to get the can method to work. Try
use IO::File;
my $fh = IO::File->new("file.out", ">>")
or die "Couldn't open file: $!";
IO::Handle doesn't overload the open() function, so you're not actually getting an IO::Handle object in $fh. I don't know why the $fh->print("Hello, world") line works (probably because you're calling the print() function, and when you do things like $foo->function it's equivalent to function $foo, so you're essentially printing to the filehandle like you'd normally expect).
If you change your code to something like:
use strict;
use IO::Handle;
open my $fh, ">>", "file.out" or die "Can't open file";
my $iofh = new IO::Handle;
$iofh->fdopen( $fh, "w" );
$iofh->print("Hello, world");
if ($iofh->can("print"))
{
print "Yes\n";
}
else
{
print "No\n";
}
...then your code will do as you expect. At least, it does for me!