make mv command quiet

[dormantgp]

I have a script to remove all extra dots in filenames:

Code: Select all

for f in *.*
do
    pre="${f%.*}"
    suf="${f##*.}"
    mv -n -i -- "$f" "${pre//./_}.${suf}"
done

If there are filenames with no extra dots, mv tells me, even when I use the -n switch.

mv: 'fred_1.txt' and 'fred_1.txt' are the same file
mv: 'fred__2.txt' and 'fred__2.txt' are the same file
mv: 'fred__3_.txt' and 'fred__3_.txt' are the same file
mv: 'fred.txt' and 'fred.txt' are the same file

Is there a way to silence mv?

[Koentje]

Code: Select all

for f in *.*
do
    pre="${f%.*}"
    suf="${f##*.}"
    mv -n -i -- "$f" "${pre//./_}.${suf}" 2>/dev/null
done

2>/dev/null redirects stderr to /dev/null
If something goes wrong, you'll never know!

[xenopeek]

Instead of making mv quiet, don't rename the file when not needed:

Code: Select all

for f in *.*
do
    pre="${f%.*}"
    suf="${f##*.}"
    if [[ "$f" != "${pre//./_}.${suf}" ]]; then
        mv -n -i -- "$f" "${pre//./_}.${suf}"
    fi
done

[Koentje]

That's indeed way better.
What i don't understand is the ${suf}? That doesn't do anything.. wouldn't it be better to use $suf instead?

[Shiva]

Is there a way to silence mv?

Pre-select files when mv would have no "objections" :

Code: Select all

find . -type f -name "*..*" | while read file; do mv "$file" "${file//../.}"; done

replaces double dots with single dots on a whole tree (find is recursively searching - subfolders too).

However commands like mv, rm.. are definitive and may prove harmful if misused. So I would first think running a dummy :

Code: Select all

find . -type f -name "*..*" | while read file; do echo mv "$file" "${file//../.}"; done

The supplemental echo outputs what will happen if the command is executed. If it's satisfactory then run the first one (without echo).

[xenopeek]

I understood the command to find files like fred.flinstone.txt and rename them to fred_flinstone.txt. Not fred..txt to fred_.txt.

[Shiva]

I understood the command to find files like fred.flinstone.txt and rename them to fred_flinstone.txt. Not fred..txt to fred_.txt.

And you're probably right. My code just replaces two consecutive dots with one.

[Shiva]

A right start for a pipeline version and the same renaming algorithm could have been :
printf "%s\n" *.* | grep '\..*\.' |

One line version :

Code: Select all

printf "%s\n" *.* | grep '\..*\.' | while read fn; do Suf=".${fn##*.}"; Pre="${fn%.*}"; echo mv "$fn" "${Pre//./_}$Suf"; done

Multi-line version with comments :

Code: Select all

printf "%s\n" *.* \
 | grep '\..*\.' \
 | while read fn; do Suf=".${fn##*.}"; Pre="${fn%.*}"; \
 echo mv "$fn" "${Pre//./_}$Suf"; done

- line 1 : prints out everything that has an extension one per line
- line 2 : keep only those having at least 2 dots : we are sure that from here on all files remaining have to be renamed according to replacing . with _ and their names will be different, so no mv error
- line 3 : while loop begin extracting Suf (extension) and Pre (prefix)
- line 4 : dummy running with echo and ending loop

If it's OK, run same command without echo.