Bash Script to read document line by line and create multiple documents with predefined filenames in same document

Question

I have a file containing a list of information looks something like this:

#FILENAME "Some name - some title.xml"    
<song>some song</song>
<year>1994.</year>
<album>Some album</album>
<artist>some artist</artist>
#FILENAME "another filename - some have ' in title title.xml"
<song>another song</song>
<year>1996.</year>
<album>another album</album>
<artist>another artist</artist>
#FILENAME "yet another filename - something.xml"
...
..
.

with more than 25 000 lines I need to create separate files (5000 files.xml) so first line is FILENAME second to fifth lines are informations that need to be fields in xml file that looks like this:

<?xml version='1.0' encoding='UTF-8'?><Metadata>    <artist></artist>    <song></song>    <album></album>    <year></year></Metadata>

Can someone help me with the script?

So far I have removed # form FILENAME from the document and done something like this:

But can't manage to create multiple files

#!/bin/bash



while read line; 
if  [[ $line == FILENAME* ]]; then
     filename="${line:9}"

fi
if [[ $line == *artist*  ]]; then
    artist=$line
fi
if [[ $line == *song* ]]; then
    song=$line
fi
if [[ $line == *album* ]]; then
    album=$line
fi
if [[ $line == *year* ]]; then
    year=$line
fi

do

    echo "<?xml version='1.0' encoding='UTF-8'?><Metadata>    $artist    $song    $album    $year</Metadata>"

done < popis.txt > $filename

Answer 1

1. You could split your file popis.txt into 5000+ temporary files with the split (GNU coreutils) command:

   split -d -a4 -l5 popis.txt split
   

   This creates the files split0001, split0002, ... with five lines per file and makes further processing a bit easier.

2. Create this modified script of yours and save it as script.sh:

   #!/bin/bash
   
   for file; do
     while read -r line; do
       if [[ "$line" = "<artist>"* ]]; then
         artist=$line
       elif [[ "$line" = "<song>"* ]]; then
         song=$line
       elif [[ "$line" = "<album>"* ]]; then
         album=$line
       elif [[ "$line" = "<year>"* ]]; then
         year=$line
       else
         # remove prefix `#FILENAME "` and the last quote `"`
         filename=$(echo "$line" | sed 's/[^"]*"//;s/"[[:space:]]*$//')
       fi
     done < "$file"
     echo "<?xml version='1.0' encoding='UTF-8'?><Metadata>${artist}${song}${album}${year}</Metadata>" > "$filename"
   done
   

3. Make your script executable and run the script on all splitXXXX files:

   chmod +x script.sh
   ./script.sh split*
   

4. If everything went well, this should have created one XML file for each input file and you can remove the temporary files:

   rm split*