script to replace spaces with -

lleb · 01-03-2008, 01:37 PM

I have seen it before but am having major failure on finding it here at LQ.

I would like a simple script that i can run in a directory that will remove all of the ' ' in a name of a file and replace them with '-' or '_' either will be fine.

also can this be run to affect sub-directories and directory names them self?

many thanks.

still ubber new to scripts and coding.

fuzzyworm · 01-03-2008, 01:54 PM

Code:

for i in *; do mv "$i" ${i// /_}; done;

should do the trick. Note that it will do this with all files and subfolders in the current directory.

Hope this helps

lleb · 01-03-2008, 01:58 PM

now for my education. as i know next to nothing about code.

i could make a bash script:

#!/bash/bin

for i in *; do mv "$i" ${i// /_};

done

what is the last ; for?

and what is the 'for i in' part of the code telling the system to do.

i get the do mv ... part as the string ($)i is everything? and the part inside of the { } is if it is // then change it to /_ or if it is a space replace the space with an _, but still not sure what the code is tell me. i see it will work, just no clue why. trying to learn.

dive · 01-03-2008, 02:06 PM

My win2lin script. It can be invoked with a file/dir name ( win2lin some\ dir ) or just on its own in which case it will change all the files/dirs in current dir. It is recursive.

Code:

#!/bin/bash

 function basic {
 for i in *
 do
     if [ -d "$i" ]
     then
         cd "$i"
         win2lin #name of this script. Must be in PATH or full path to script must be given.
         cd ..
     fi
     tempa=$(echo "$i" | sed 's/ /_/g' | sed 's/_-_/-/g' | sed 's/\ -\ /-/g' | sed 's/__/_/g' \
     | sed "s#'##g" | sed "s#;##g" | sed "s#,##g")
     [[ "$i" != "$tempa" ]] && mv -v "$i" "$tempa"
 done
 }

 function cmd {
 for i in "$@"
 do
     tempa=$(echo "$i"  | sed 's/ /_/g' | sed 's/_-_/-/g' | sed 's/\ -\ /-/g' | sed 's/__/_/g' \
     | sed "s#'##g" | sed "s#;##g" | sed "s#,##g")
     [[ "$i" != "$tempa" ]] && mv -v "$i" "$tempa"
     [[ ! -d "$tempa" ]] && break
     cd "$tempa"
     win2lin #name of this script. Must be in PATH or full path to script must be given.
     cd ..
 done
 }

 [[ "$@" != "" ]] && cmd "$@"
 [[ "$@" == "" ]] && basic

fuzzyworm · 01-03-2008, 02:14 PM

That script ought to cover it better than my one line command.

However, in answer to your question:

a) I like to finish my lines with ';', but it should work without the final one.

b) (See man page for further info: http://www.ss64.com/bash/for.html )

Code:

for i in *; do ...; done;

Takes * (all files in current folder [or *.*, in some cases])
With each one in turn, sets it to 'i', then executes the code

Hope this has helped.

lleb · 01-03-2008, 02:16 PM

ok now i understand the for i in * part. thank you for clearing that up.

lleb · 01-03-2008, 02:34 PM

Quote:

Originally Posted by dive

My win2lin script. It can be invoked with a file/dir name ( win2lin some\ dir ) or just on its own in which case it will change all the files/dirs in current dir. It is recursive.

Code:

#!/bin/bash

 function basic {
 for i in *
 do
     if [ -d "$i" ]
     then
         cd "$i"
         win2lin #name of this script. Must be in PATH or full path to script must be given.
         cd ..
     fi
     tempa=$(echo "$i" | sed 's/ /_/g' | sed 's/_-_/-/g' | sed 's/\ -\ /-/g' | sed 's/__/_/g' \
     | sed "s#'##g" | sed "s#;##g" | sed "s#,##g")
     [[ "$i" != "$tempa" ]] && mv -v "$i" "$tempa"
 done
 }

 function cmd {
 for i in "$@"
 do
     tempa=$(echo "$i"  | sed 's/ /_/g' | sed 's/_-_/-/g' | sed 's/\ -\ /-/g' | sed 's/__/_/g' \
     | sed "s#'##g" | sed "s#;##g" | sed "s#,##g")
     [[ "$i" != "$tempa" ]] && mv -v "$i" "$tempa"
     [[ ! -d "$tempa" ]] && break
     cd "$tempa"
     win2lin #name of this script. Must be in PATH or full path to script must be given.
     cd ..
 done
 }

 [[ "$@" != "" ]] && cmd "$@"
 [[ "$@" == "" ]] && basic

now that is a might set of code for my little head to wrap around. mind educating me a bit. what is sed?

i can see that sed is looking at a situation of the ls output and if it finds something it is going to replace it with the other... so sed 's/ /_/g' if i understand is telling me that string 's' of / to replace with _ and the /g is go? kind of like in vi editing?

what is the rest of it tell me to do?

lleb · 01-03-2008, 02:39 PM

Code:

sh remove_space2.sh
remove_space2.sh: line 32: unexpected EOF while looking for matching `"'
remove_space2.sh: line 33: syntax error: unexpected end of file

now when i look at the code i typed in with vi :set number i only show 32 lines of code...

Code:

1 #!/bin/bash
 2 #
 3 #       remove spaces in names of files and replace them with '_'
 4 #
 5         function basic {
 6         for i in *
 7         do
 8                 if [ -d "$i$ ]
 9                 then
10                         cd "$i"
11                         win2lin #name of this script from LQ member dive.  Must be in PATH or full path t   o script must be given.
12                         cd ..
13                 fi
14                 tempa=$(echo "$i" | sed 's/ /_/g' | sed 's/_-_/-/g' | sed -e 's/\ -\ /-/g' | sed 's/__/_/   g' \ | sed "s#'##g" | sed "s#;##g' | sed "s#,##g")
15                 [[ "$i" != "$tempa" ]] && mv -v "$i" "$tempa"
16         done
17         }
18
19         function cmd {
20         for i in "$@"
21         do
22                 tempa=$(echo "$i" | sed 's/ /_/g' | sed 's/_-_/-/g' | sed -e 's/\ -\ /-/g' | sed 's/__/_/   g' \ | sed "s#'##g" | sed "s#;##g" | sed "s#,##g")
23                 [[ "$i" != "$tempa" ]] && mv -v "$i" "$tempa"
24                 [[ ! -d "$tempa" ]] && break
25                 cd "$tempa"
26                 win2lin #name of this script from LQ member dive.  Must be in PATH or full path to script    must be given.
27                 cd ..
28         done
29         }
30
31         [[ "$@" != "" ]] && cmd "$@"
32         [[ "$@" == "" ]] && basic
:set number

so were did i screw up

dive · 01-03-2008, 03:26 PM

I can see line 8 is wrong and causes that error. Also 11 and 26 need to point to this script. If you rename it to remove_space2.sh then replace win2lin with that and either put it in your path or provide full path to the script.

14 and 22 are also wrong. You have made 2 lines into 1. The \ at the end of each line tells bash that the command continues on the next line.

dive · 01-03-2008, 03:29 PM

sed = stream editor. In

sed 's/ /_/g'

The s means substitute. The g means global. So the command tells sed to substitute all spaces ( / / ) with underscore ( /_/ ) globally.

There are other sed commands in there to make __ into _ and also remove ; and ,

lleb · 01-03-2008, 04:42 PM

ok made these changes...

1. adjusted line 8, 13, 22 (they are now 8, 14, and 23) to fix typo and add a break (hit the return key)
2. changed name from win2lin to remove_space2.sh as that is what i created it as, that should work i hope.
3. still getting errors... what exactly do you mean put it in my path?

can i just be in the directory of the script and do : sh remove_space2.sh and have it run on that directory?

Code:

sh remove_space2.sh 
remove_space2.sh: line 34: unexpected EOF while looking for matching `"'
remove_space2.sh: line 35: syntax error: unexpected end of file
ray@sabAMD64 ~/torrents $ cat remove_space2.sh 
#!/bin/bash
#
#       remove spaces in names of files and replace them with '_'
#
        function basic {
        for i in *
        do
                if [ -d "$i" ]
                then
                        cd "$i"
                        remove_space2.sh #name of this script from LQ member dive.  Must be in PATH or full path to script must be given.
                        cd ..
                fi
                tempa=$(echo "$i" | sed 's/ /_/g' | sed 's/_-_/-/g' | sed -e 's/\ -\ /-/g' | sed 's/__/_/g' \
                | sed "s#'##g" | sed "s#;##g' | sed "s#,##g")
                [[ "$i" != "$tempa" ]] && mv -v "$i" "$tempa"
        done
        }

        function cmd {
        for i in "$@"
        do
                tempa=$(echo "$i" | sed 's/ /_/g' | sed 's/_-_/-/g' | sed -e 's/\ -\ /-/g' | sed 's/__/_/g' \
                | sed "s#'##g" | sed "s#;##g" | sed "s#,##g")
                [[ "$i" != "$tempa" ]] && mv -v "$i" "$tempa"
                [[ ! -d "$tempa" ]] && break
                cd "$tempa"
                remove_space2.sh #name of this script from LQ member dive.  Must be in PATH or full path to script must be given.
                cd ..
        done
        }

        [[ "$@" != "" ]] && cmd "$@"
        [[ "$@" == "" ]] && basic

so i hope i made the fixes you caught correctly.

also thank you very much for explaining what sed is and the s and g. to me this is one of the FOSS communities greatest strengths. the willingness of people to take time to educate others.

dive · 01-03-2008, 05:10 PM

Quote:

Originally Posted by lleb

Code:

| sed "s#'##g" | sed "s#;##g' | sed "s#,##g")

There's an error here: the 2nd sed command starts with a " and ends with a '.

Should be:

Code:

| sed "s#'##g" | sed "s#;##g" | sed "s#,##g")

Are you typing all this manually? Easier to copy and paste.

For the lines that call the script again you will probably need to put /path/to/remove.space2.sh

But unless this is the only script that you ever want to use it would be eaiser in the long run to add a scripts folder in your home directory and then add it to your PATH env variable in .bashrc or .bash_profile:

export PATH=$PATH:~/scripts

This way you will only ever need to call the name of the script.

ghostdog74 · 01-03-2008, 06:08 PM

Quote:

Originally Posted by lleb

I have seen it before but am having major failure on finding it here at LQ.

I would like a simple script that i can run in a directory that will remove all of the ' ' in a name of a file and replace them with '-' or '_' either will be fine.

Code:

   for files in * 
    do
        if [ -f "$files" ];then
            case $files in 
            *" "* ) 
                    newfile=`echo $files | sed 's/ /./g'`
                    mv "$files" "$newfile"
                    ;;    
            esac
        fi
    done

Quote:

also can this be run to affect sub-directories and directory names them self?

yes. I leave it to you to try it out yourself.

Quote:

many thanks.

Thank this guy instead

Quote:

still ubber new to scripts and coding.

Then you should his advice first.

David the H. · 01-03-2008, 06:31 PM

It's a little late here, but I always use the perl rename (aka prename) command, a script provided as part of the perl package. It uses the basic sed replacement syntax.

rename -v "y/\ /_/" ./*

To run it on subdirectories simply use the find command:

find . -exec rename -v "y/\ /_/" '{}' \;

I've been using it all weekend to clean up the file names on my music directory. I've even created an alias from it specifically for quickly removing spaces from filenames.

lleb · 01-03-2008, 11:10 PM

Quote:

Originally Posted by dive

There's an error here: the 2nd sed command starts with a " and ends with a '.

Should be:

Code:

| sed "s#'##g" | sed "s#;##g" | sed "s#,##g")

Are you typing all this manually? Easier to copy and paste.

Yes, when i tried to copy/paste i ended up with lines that were way out of wack so i just did it manually... thus all the typo's

Quote:

For the lines that call the script again you will probably need to put /path/to/remove.space2.sh

But unless this is the only script that you ever want to use it would be eaiser in the long run to add a scripts folder in your home directory and then add it to your PATH env variable in .bashrc or .bash_profile:

export PATH=$PATH:~/scripts

This way you will only ever need to call the name of the script.

ok still not 100% on how to add things to my .bashrc as in my home directory the only .bashXXX i have is my bash history. gentoo/sabayon does some strange things when it comes to that. i had a devil of a time getting an alias to be part of the bash environment.

once i replaced the ' with " it worked like a charm. thank you.