Delete sub-directories based on matching information from column

Question

I want to keep only those sub-directories that are present in the corresponding directory from CSV file. The file structure looks something like this:

100_folder/
├── folder_11
├── folder_25
├── folder_31
└── folder_41
210_folder/
├── folder_13
├── folder_23
├── folder_31
└── folder_42

Information in CSV:

Col6,Col26
100,folder_11
100,folder_13
100,folder_41
210,folder_31
210,folder_42

Based on the information from the columns, I want to remove the sub-directories that are not present in CSV file.

Here is how I read the file:

eCollection=( $(cut -d ',' -f6,26 file.csv ) )
echo "${eCollection[@]}"

Answer 1

Of course we can provide solution for you. But where is the fun?

Let me say your requirements look dangerous to me, as every folder not in csv will deleted (imagine typo, wrong file format/line ending or trailing whitespaces)

That said, I will introduce three friends of text file processing in shell (script)

• string manipulation (of variables)
• grep text search core-util (file content, read only)
• find file search core-util (file name)

Never use code you don't fully understand!

#!/bin/sh

csvfile='index.txt'
csvseparator=','

cut -d ',' -f6,26 file.csv > index.txt

for subdir in ./*/*/
  do
    subdir=${subdir%/}
    dir=${subdir%/*}
    parent=${dir%/*}
    subdir=${subdir##*/}
    dir=${dir##*/}
    if grep -Fxq "${dir%_*}$csvseparator$subdir" "$csvfile"
      then
        echo "ok: $parent/$dir/$subdir"
      elif grep -wq "^${dir%_*}" "$csvfile"
        then
          echo "no: $parent/$dir/$subdir"
#          find "$parent/$dir/$subdir" -delete
    fi
done