I have a perl script which takes a fasta file as input. So from command line to run this I have to execute this: "perl scripl.pl some_seq.fasta some_directory". But, I have to do this for hundreds of fasta file. Is there any way using bash script to automate this process?
Using wildcards to select a set of files , and using the bash for statement, you can issue (even from the command line)
If more changes than just the filename, you may need to be more creative
with manipulating bash variables,
for f in *.fast ; do
perl scripl.pl $f some_directory
done
Using wildcards to select a set of files , and using the bash for statement, you can issue (even from the command line)
If more changes than just the filename, you may need to be more creative
with manipulating bash variables,
for f in *.fast ; do
perl scripl.pl $f some_directory
done
Hi,
you can also use another perl script. In example,
##############
use warnings;
use strict;
my ($working_dir) = @ARGV; #This will catch the desired directory in $working_dir #from the command line
#Place the path of all fasta files in @files array
my @files = glob("$working_dir/*.fasta");
foreach my $file (@files){
#Do a system call for your script.
system "perl scripl.pl $file some_directory";
}
#############
This is a simple example. I generally prefer perl over bash since you can perform more complex duties easier. But that is just a matter of taste.
hi,
firstly, if you put
#!/usr/bin/perlin the first line of the script and you make it executable (chmod a+x scriptname), you need not put "perl" in front of the script, just use
scripl.pl some_seq.fasta some_directory
secondly, if you want bash solution, try this (you have *.fasta xargs -n 1 files and you want to treat them all of them with somedirectory and you have not inserted the above line into the script yet):
find . -name "*.fasta" -exec perl scriptname {} somedirectory \;
note the escaped semicolon at the end. In other cases if the file name goes to the last parameters, it would be easier to use:
ls -1 "*.fasta" | scriptname parameters
or
ls -1 | grep fasta | xargs -n 1 scriptname parameters
but only find ... -exec provides the possibility to put the file name into whichever position of the argument list using {}
third,
if you have hundreds of files, it MAY be OK. If you have thousands, it will not work, as the shell has some size limitation for the list of all the command invocations, which list is created somewhere before actually starting your script for the first time.
Then you need some other solutions. I also suggest using a perl script. I write here the same as above in another way, just for fun. :)
#!/usr/bin/perl -w
opendir (DIR, '.') or die "Couldn't open directory, $!";
$directory = $ARGV[0] or die "\n usage: $0 some_directory\n";
while ($file = readdir DIR) {
next unless ($file =~ /\.fasta\z/);
system ("perl scriptname $file $directory");
}
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology