在 bash 中创建文件的备份

Question

I want to write into a file in a bash script but I want to make sure that the file is backed up if it exists and I also want to avoid overwriting any existing backups.

So basically I have $FILE, if this exists, I want to move $FILE to $FILE.bak if it does not already exist, otherwise to $FILE.bak2, $FILE.bak3, etc.

Is there a shell command for this?

Answer 1

Using a function to find the next available name:

#!/usr/bin/env bash
function nextsuffix {
    local name="$1.bak"
    if [ -e "$name" ]; then
        printf "%s" "$name"
    else
        local -i num=2
        while [ -e "$name$num" ]; do
            num+=1
        done
        printf "%s%d" "$name" "$num"
    fi
}

mv "$1" "$(nextsuffix "$1")"

If foo.bak already exists, it just loops until a given foo.bakN filename doesn't exist, incrementing N each time.

Answer 2

You can just output to a file with a date.

FILE=~/test
echo "123" >> $FILE.$(date +'%Y%d%m')

If you want the numbers logrotate seems to be most ideal.

Answer 3

cp "$FILE" "$FILE.bak$(( $(grep -Eo '[[:digit:]]+' <(sort -n <(for fil in $FILE.bak*;do echo $fil;done) | tail -1 )) + 1 ))"

Breaking the commands down

sort -n <(for fil in $FILE.bak*;do echo $fil;done) | tail -1

List the last file in the directory which is sorted in numeric form

grep -Eo '[[:digit:]]+' <(sort -n <(for fil in $FILE.bak*;do echo $fil;done) | tail -1 ))

Strip out everything but the digits

(( $(grep -Eo '[[:digit:]]+' <(sort -n <(for fil in $FILE.bak*;do echo $fil;done) | tail -1 )) + 1 ))

Add one to the result

Answer 4

For posterity, my function with changes inspired by @Shawn's answer

backup() {
    local file new n=0
    local fmt='%s.%(%Y%m%d)T_%02d'
    for file; do
        while :; do
            printf -v new "$fmt" "$file" -1 $((++n))
            [[ -e $new ]] || break
        done
        command cp -vp "$file" "$new"
    done
}

I like to cp not mv .