用于删除旧备份文件的现有 bash 脚本 - 它是如何工作的？

Question

I'm a total beginner with scripts. I have some questings regarding an old script, which should delete old backup files.

KEEP_FULL=7
KEEP_DAILY=14
KEEP_WEEKLY=30

DB_FULL_PATH=/Volumes/path

LAST_DAY=0
LAST_WEEK=0
LAST_MONTH=0
find $DB_FULL_PATH -type f| while read f; do
    < <(stat -f %Sm -t "%m %V %d" $f) read -s MONTH WEEK DAY 
    if [ $DAY -eq $LAST_DAY ]; then
        find $f -mtime +$KEEP_FULL | xargs rm 
    else if [ $WEEK -eq $LAST_WEEK ]; then
        find $f -mtime +$KEEP_DAILY | xargs rm 
    else if [ $MONTH -eq $LAST_MONTH ]; then
        find $f -mtime +$KEEP_WEEKLY | xargs rm 
    fi fi fi
    export LAST_DAY=$DAY
    export LAST_WEEK=$WEEK
    export LAST_MONTH=$MONTH
done

Could someone explain (for dummies) what happens within the while-loop? I understand that for each file within the folder the information (day, week, month of creation) is written into $MONTH $WEEK $DAY. But the following logic I don't understand correctly.

Answer 1

This answer won't be complete because I lack informations. Also It seems overly complex for the task it should do.

---

The loop below process all files under /Volumes/path and stores them in the variable f .

find $DB_FULL_PATH -type f| while read f; do
    # [...]
done

For each file f , this command is first performed:

< <(stat -f %Sm -t "%m %V %d" $f) read -s MONTH WEEK DAY 

1. stat is invoked on the file f
   • According to the man page of stat (OSX):

   -f format Display information using the specified format. See the FORMATS section for a description of valid formats.

   -t timefmt Display timestamps using the specified format. This format is passed directly to strftime (3).

   • According to the FORMATS section of the manual page, %Sm is the combination of S and m , which stand respectively for 1) "the [last] time file was [...] modified " ( m ) and 2) formatted, "if applicable, should be in string format" ( S , it should be the case here).
   • According to the manual page of strftime , %m %V %d stand for the month , the week number of the year and the day of the month as a decimal number
2. The output is read and stored in variables MONTH WEEK and DAY ( -s stands for silent mode )

At this point, you have the month , the week number of the year and the day of the month of the time the file was modifed.

Next the hardest part:

• If the day of modification of the previously processed file equals the current processed file one, then check if it was modified ( -mtime ) one week ago ( KEEP_FULL =7) ; if so, delete it ( f is passed to rm with xargs ):

   if [ $DAY -eq $LAST_DAY ]; then find $f -mtime +$KEEP_FULL | xargs rm fi

• Otherwise, if the week number of the year of modification of the previously processed file ( breathes ) equals the current processed file one, then check if it was modified two weeks ago ( KEEP_DAILY =14) ; if so, delete it:

   else if [ $WEEK -eq $LAST_WEEK ]; then find $f -mtime +$KEEP_DAILY | xargs rm

• Otherwise, if the month of modification of the previously processed file ( sips ) equals the current processed file one, then check if it was modified 30 days ago ( KEEP_WEEKLY =30) ; if so, delete it:

   else if [ $MONTH -eq $LAST_MONTH ]; then find $f -mtime +$KEEP_WEEKLY | xargs rm fi fi fi

• Retrieve the month, the week and the day of the month of the current file to process the next one with them (will be used as parts of the modification time of said "previously processed" file):

   export LAST_DAY=$DAY export LAST_WEEK=$WEEK export LAST_MONTH=$MONTH done

That said, to understand why such comparisons are performed, you need to know the order the first find outputs files injected to the standard input of the while loop.