Linux Rename

Jhinukk · 03-28-2013, 05:33 AM

Hi,

I have 2 files. Which looks like

1st file -------------- 2nd file
-------------- ----------------
ABC_old.txt ------------ ABC_new.txt
XYZ_old.txt ------------ XYZ_new.txt

Now I have to use 'mv' command using this files in my script such as it will come like,

mv ABC_old.txt ABC_new.txt
mv XYZ_old.txt XYZ_new.txt

TIA,
Jhinukk

Ramurd · 03-28-2013, 05:40 AM

and what exactly is the function of all those dashes? Are they always the same amount of dashes?

If the amount of dashes differs, you could approach it like this

Code:

while read line
do
OLDNAME=$(echo $line | cut -d ' ' -f 1)
NEWNAME=$(echo $line | rev | cut -d ' ' -f 1 | rev)
mv ${OLDNAME} ${NEWNAME}
done

This is provided that there are never spaces in the filenames and is pretty quick and dirty.

Jhinukk · 03-28-2013, 06:16 AM

Thanks Ramurd

. It solved my problem.

David the H. · 03-28-2013, 06:47 AM

1) Please use ***[code][/code]*** tags around your code and data, to preserve the original formatting and to improve readability. Do not use quote tags, bolding, colors, "start/end" lines, or other creative techniques.

2) There have been many, many previous threads on how to bulk rename files in the shell. A bit of searching through the archives would certainly have turned up the answers you needed. We'd be here afterwards if you still needed help.

But for my own suggestion, I'm assuming that the OP just has two files, with each line in the first file having the old name, and with the second file holding a corresponding new name.

For that I'd recommend a couple of simple bash arrays:

Code:

mapfile -t oldname <file1.txt
mapfile -t newname <file2.txt

for i in "${!oldname[@]}"; do
    echo mv "${oldname[i]}" "${newname[i]}"
done

It also assumes one line per name pair, with none of the names containing newlines themselves, and that the filenames either exist in the current directory or have absolute paths.

I added an echo to the front of mv so you can confirm the commands before actually executing the changes. Just remove it when ready.

Ramurd · 03-28-2013, 06:48 AM

You're welcome;

Please mark the thread as solved if it did solve your issue...

And to make things a bit easier, I can explain what the snippet does:

Code:

while read line
do
done < ${INPUTFILE}

This part walks throug the file, line by line.

Code:

OLDNAME=$(echo $line | cut -d ' ' -f 1)

Split the line on spaces and take the first word; The weakness is that if the filename contains a space, the cutting is done on that space.

Code:

NEWNAME=$(echo $line | rev | cut -d ' ' -f 1 | rev)

This one takes the same line, but prints it in reverse (rev), cuts it on a space (again) and takes the first word (since the line is in reverse, this is actually the last "word"); Of course, we have to reverse the reverse, so things are in the correct order: that's what the second "rev" does.

David the H. · 03-28-2013, 07:28 AM

Well that's a lot of trouble. Why not just use the shell's built-in parameter substitutions?

Code:

oldname=${line%% *}
newname=${line##* }

Another option is to use read itself (assuming bash's array support).

Code:

While read -ra line; do

     oldname=${line[0]}
     newname=${line[-1]}
     <mv...>

done <infile.txt

But then again these are assuming a single file, with the first and last "words" being the filenames, which is not what the OP described.

mina86 · 03-28-2013, 04:44 PM

Many distributions have a rename command. Beware that some have a version which uses perl expressions and others have a version that uses old and new parameters. With the former, you could do:

Code:

rename 's/_old.txt/_new.txt/' *_old.txt

with the latter you could do:

Code:

rename _old.txt _new.txt *_old.txt