无法用Sed / Python / Perl替换给定文件夹内容中的单词

Question

I have a project where I have folders, subfolders, and files. I need to replace the word Masi by the word Bond in each files.

I run the following Sed script called replace unsuccessfully

s/Masi/Bond/

in Zsh by

sed -f PATH/replace PATH2/project/**

It gives me all files, also the ones which do not have Masi, as an output.

Sed is not necessarily the best tool for the task. I am interested in Python and Perl.

How would you do the replacement in Sed/Perl/Python, such that only the file contents are changed?

Answer 1

To replace the word in all files found in the current directory and subdirectories

perl -p -i -e 's/Masi/Bond/g' $(grep -rl Masi *)

The above won't work if you have spaces in filenames. Safer to do:

find . -type f -exec perl -p -i -e 's/Masi/Bond/g' {} \;

or in Mac which has spaces in filenames

find . -type f -print0 | xargs -0 perl -p -i -e 's/Masi/Bond/g'

Explanations

• -p means print or die
• -i means "do not make any backup files"
• -e allows you to run perl code in command line

Answer 2

Why not just pass the -i option ( man sed ) to sed and be done with it? If it doesn't find Masi in a file, the file will just be rewritten with no modification. Or am I missing something?

If you don't want to replace the files' contents inline (which is what the -i will do) you can do exactly as you are now, but throw a grep & xargs in front of it:

grep -rl Masi PATH/project/* | xargs sed -f PATH/replace

Lots of options, but do not write an entire perl script for this (I'll give the one-liner a pass ;)). find , grep , sed , xargs , etc. will always be more flexible, IMHO.

In response to comment:

grep -rl Masi PATH/project/* | xargs sed -n -e '/Masi/ p'

Answer 3

Renaming a folder full of files:

use warnings;
use strict;
use File::Find::Rule;

my @list = File::Find::Rule->new()->name(qr/Masi/)->file->in('./');

for( @list ){
   my $old = $_;
   my $new = $_;
   $new =~ s/Masi/Bond/g;
   rename $old , $new ;
}

Replacing Strings in Files

use warnings;
use strict;
use File::Find::Rule;
use File::Slurp;
use File::Copy;

my @list = File::Find::Rule->new()->name("*.something")->file->grep(qr/Masi/)->in('./');

for( @list ){
   my $c = read_file( $_ );
   if ( $c =~ s/Masi/Bond/g; ){
    File::Copy::copy($_, "$_.bak"); # backup.
    write_file( $_ , $c );
   }
}

• strict (core) - Perl pragma to restrict unsafe constructs
• warnings (core) - Perl pragma to control optional warnings
• File::Find::Rule - Alternative interface to File::Find
• File::Find (core) - Traverse a directory tree.
• File::Slurp - Efficient Reading/Writing of Complete Files
• File::Copy (core) - Copy files or filehandles

Answer 4

A solution tested on Windows

Requires CPAN module File::Slurp. Will work with standard Unix shell wildcards. Like ./replace.pl PATH/replace.txt PATH2/replace*

#!/usr/bin/perl

use strict;
use warnings;
use File::Glob ':glob';
use File::Slurp;
foreach my $dir (@ARGV) {
  my @filelist = bsd_glob($dir);
  foreach my $file (@filelist) {
    next if -d $file;
    my $c=read_file($file);
    if ($c=~s/Masi/Bond/g) {
      print "replaced in $file\n";
      write_file($file,$c);
    } else {
      print "no match in $file\n";
    }
  }
}

Answer 5

import glob
import os

# Change the glob for different filename matching 
for filename in glob.glob("*"):
  dst=filename.replace("Masi","Bond")
  os.rename(filename, dst)