Script to split string

Question

@terdon's answer provides this script:

find . -type f -name "*mp3" | while read file; do 
    album="$(basename "$(dirname "$file")")"; 
    filename="$(basename "$file")"; 
    artist=${filename%%-*}; 
    title=${filename##*-}; 
    title=${title%%.mp3}; 
    eyeD3 -A "$album" -t "$title" -a "$artist" "$file"; 
done

It makes use of eyeD3 to update mp3 tags. It extracts Album, Artist and Title from file name, supposing that the file name is Album/Artist - Title.mp3.

Well, after running it the Title field is starting with a blank space. See below:

How to fix this if my mp3 files have the pattern

Album/<Artist with possible spaces> - <track number> - <Title with possible spaces>.mp3

Answer 1

Since I assume that your names do not always have these spaces, the easiest thing to do would be to simply remove the space if present:

find . -type f -name "*mp3" | 
  while read file; do 
   album="$(basename "$(dirname "$file")")"; 
   filename="$(basename "$file")"; 
   artist=${filename%%-*}; artist=${artist%% }; artist=${artist## };
   title=${filename##*-}; title=${title%% }; title=${title## };
   title=${title%%.mp3}; 
   eyeD3 -A "$album" -t "$title" -a "$artist" "$file";
 done

Answer 2

Ksh extended glob patterns give wildcards the power of regular expressions. In bash, turn them on with shopt -s extglob.

Then, instead of

title=${filename##*-}

to remove the part of $filename up to the last -, you can use

title=${filename##*- #}

to also strip whitespace after the -.

Answer 3

If you just want to remove the leading space for the Artist, you could use sed. Something like this should work.

artist=${filename%%-*}; 
artist=`echo $artist |  sed 's/^ *//g'`;