[SOLVED] Renaming a set of files.

stf92 · 03-04-2014, 03:54 PM

Hi: I have a set of files under the same directory, all filenames beginning with the same prefix. How could I take away the prefix from all the filenames?

All I need is a way to do the operation for only one file, as then I could write a script to do the same within a loop.

stf92 · 03-04-2014, 04:12 PM

OK. Sed is the answer.

Richard Cranium · 03-04-2014, 10:54 PM

See "man rename"

kikinovak · 03-05-2014, 04:21 AM

Quote:

Originally Posted by stf92

Hi: I have a set of files under the same directory, all filenames beginning with the same prefix. How could I take away the prefix from all the filenames?

All I need is a way to do the operation for only one file, as then I could write a script to do the same within a loop.

If you use the Xfce desktop, there's a nifty graphical application for just what you are trying to do: Thunar Bulk Rename. I use it all the time for mass-renaming music files for my car stereo's USB stick, which can't handle subdirectories.

eloi · 03-05-2014, 04:44 AM

I do not remember where I found this perl script (sorry to the author):

Code:

#!/usr/bin/perl

($regexp = shift @ARGV) || die "Usage:  rename perlexpr [filenames]\n";

if (!@ARGV) {
	@ARGV = <STDIN>;
	chomp(@ARGV);
}


foreach $_ (@ARGV) {
	$old_name = $_;
	eval $regexp;
	die $@ if $@;
	rename($old_name, $_) unless $old_name eq $_;
}

exit(0);

This script is more complete (it allows you to use regex) than the rename
version that comes with Slackware.

(Put it in ~/bin and rename it to i.e. prename)

lems · 03-05-2014, 09:41 AM

If in the future you might need to rename files again, you might also want to check out rename: a perl script by ex-Slackware user Aristotle Pagaltzis. There's also a tutorial.

rkfb · 03-05-2014, 03:30 PM

Quote:

Originally Posted by lems

If in the future you might need to rename files again, you might also want to check out rename: a perl script by ex-Slackware user Aristotle Pagaltzis. There's also a tutorial.

Although you might want to rename it first to avoid conflict with the stock rename in /usr/bin

stf92 · 03-06-2014, 09:30 AM

Thanks for suggesting 'rename'. But could somebody tell me what's the problem with this script?:

Code:

semoi@server:~/STORE1/Nonsoft/libros/chess/nuevo$ cat ren.sh
#!/bin/bash

# PROG APPL: Rename a set of files.

# El script es poco flexible. No tiene argumentos de entrada. It's hard
# coded. In current dir there are
# Averbakh,\ Yuri\ -\ Comprehensive\ 1.djvu
# Averbakh,\ Yuri\ -\ Comprehensive\ 2.djvu
# Averbakh,\ Yuri\ -\ Comprehensive\ 3.djvu

# The script renames them to
# BOOK01
# BOOK02
# BOOK03
# s is sed's substitute command, like s in vim.

LIST="$(ls Aver*)" #ls  interpretado como orden por los parentesis
for i in "$LIST" ; do
        INPUT="$i"
        OUTPUT=$(echo "$INPUT" | tr ' ' '_')
#        OUTPUT=$(echo "$OUTPUT" | sed s/Averbakh,_Yuri_-_Comprehensive_//)
#        mv "$INPUT" "$OUTPUT"	
        echo -n "Output= "
        echo "$OUTPUT"
done
# The second command could be tr as well, or both commands sed. Either
# tr or sed can be used in both lines.

#Averbakh,\ Yuri\ -\ Comprehensive\ Chess\ Endings\ 1\ -\ Bishop\ &\ Knight\ Endings.pdf*

semoi@server:~/STORE1/Nonsoft/libros/chess/nuevo$

It should replace spaces with underscores ('_'). OUTPUT should contain a single filename. However, the loop is traversed only once, which makes me think OUTPUT contains all of the filenames concatenated:

Code:

semoi@server:~/STORE1/Nonsoft/libros/chess/nuevo$ ./ren.sh
Output= Averbakh,_Yuri_-_Comprehensive_Chess_Endings_1_-_Bishop_&_Knight_Endings.pdf
Averbakh,_Yuri_-_Comprehensive_Chess_Endings_1_-_Bishop_&_Knight.djvu
Averbakh,_Yuri_-_Comprehensive_Chess_Endings_2_-_Bishop_vs_Knight,_Rook_vs_Minor_Pieces.pdf
Averbakh,_Yuri_-_Comprehensive_Chess_Endings_3_-_Queen_Endings.djvu
Averbakh,_Yuri_-_Comprehensive_Chess_Endings_3_-_Queen_Endings.txt
Averbakh,_Yuri_-_Comprehensive_Chess_Endings_3_-_Queen_Endings.txt~
Averbakh,_Yuri_-_Comprehensive_Chess_Endings_4_-_Pawn_Endings.djvu
Averbakh,_Yuri_-_Comprehensive_Chess_Endings_4_-_Pawn_Endings.pdf
Averbakh,_Yuri_-_Comprehensive_Chess_Endings_5_-_Rook_Endings.pdf
semoi@server:~/STORE1/Nonsoft/libros/chess/nuevo$

As you can see, the constant "Output= " is output only once!

Richard Cranium · 03-06-2014, 09:34 AM

Your loop is only looping once. Put an "echo $i" at the beginning of the loop to verify.

colucix · 03-06-2014, 11:21 AM

An aside note: the lines

Code:

LIST="$(ls Aver*)"
for i in $LIST ; do

can be condensed into

Code:

for i in Aver* ; do

eloi · 03-07-2014, 03:28 AM

Code:

for i in Aver* ; do
	mv "$i" `echo "$i" | tr ' ' '_'`
done

stf92 · 03-07-2014, 04:36 AM

Quote:

Originally Posted by colucix

An aside note: the lines

Code:

LIST="$(ls Aver*)"
for i in $LIST ; do

can be condensed into

Code:

for i in Aver* ; do

Writing it in the condensed form solves the problem. I mean, now the loop is traveled as many times as there are files in the directory. But now I want to know what the trouble is with

Code:

LIST="$(ls Aver*)"
for i in "$LIST" ; do

Why is the loop traveled only once? If I write $LIST without quotes then the spaces in the filenames make the loop to be traveled once for every word in the filename.

colucix · 03-07-2014, 05:07 AM

Quote:

Originally Posted by stf92

Why is the loop traveled only once? If I write $LIST without quotes then the spaces in the filenames make the loop to be traveled once for every word in the filename.

If you use quotes the whole list is taken as a single word! To avoid problems with blank spaces, either change the IFS variable or use a while read loop, redirecting standard input from <(find Aver*).

55020 · 03-07-2014, 05:16 AM

Quote:

Originally Posted by stf92

Code:

LIST="$(ls Aver*)"
for i in "$LIST" ; do

Why is the loop traveled only once? If I write $LIST without quotes then the spaces in the filenames make the loop to be traveled once for every word in the filename.

When you store the list of filenames in a variable, there is no way of telling the difference between (1) spaces that were inside filenames, and (2) spaces that were between filenames. They are indistinguishable. A space is a space, it does not have a homoeopathic memory of whether it used to be inside a filename or between two filenames. You can either quote the whole lot as a single string, or not quote it as a list of words. There is no other choice.

The usual ways out of this problem are (a) don't allow spaces in your filenames, (b) use shell filename expansion like eloi showed you, (c) use an array (bash-specific, read the manual), (d) if you are writing a script, set the positional parameters (set -- Aver*) and use them as "$@", (e) use the find command with '-print0' and the xargs command with --null (read the manual for details).

If you are now feeling a bit sick, this is why most people choose option (a).

Note also that if there are *no* matching files, shell filename expansion of 'Aver*' will return the string 'Aver*' and the loop will be executed once with i='Aver*' (which does not exist). This might give the user a helpful error message, or it might cause an unhelpful error. To avoid this, test the value of i within the loop and do nothing if it equals 'Aver*', or set the shell option 'nullglob'.

55020 · 03-07-2014, 05:25 AM

Quote:

Originally Posted by colucix

If you use quotes the whole list is taken as a single word! To avoid problems with blank spaces, either change the IFS variable or use a while read loop, redirecting standard input from <(find Aver*).

while read has its own problems -- it will strip any spaces from the start of each filename and collapse multiple spaces inside filenames

IFS is a game you can't win. Choose any random character that you *can't* have in your filenames. Are you feeling lucky?