在Shell中使用正则表达式查找命令以查找具有两个扩展名的文件

Question

I am trying to list generated log and zip files from my application server.

• Files which are .log or .zip
• These files include digits in their name. ie Files with any number of digits in their name
• Files should be older than +5 days.

I used below expression. but looks something wrong. Could you please assist with regular expression?

ROOT_DIR=applications/jboss-as/servers/
find $ROOT_DIR -name '*[0-9]*[zip|log]' -mtime +5

Finally I wish to delete these files using command

find $ROOT_DIR -name '*[0-9]*[zip|log]' -mtime +5 -exec rm {} \;

Answer 1

The first command will find them and display.

find $ROOT_DIR ! -readable -prune -mtime +5 -type f |  egrep -e "^.*\.(log|zip)$"

The second one will remove them all

find $ROOT_DIR ! -readable -prune -mtime +5 -type f |  egrep -e "^.*\.(log|zip)$" | xargs -L 1 rm

Answer 2

You could do it this way (with most versions of find ):

find "$ROOT_DIR" '(' -name '*[0-9]*.log' -o -name '*[0-9]*.zip' ')' -mtime +5 -exec rm {} +

The + is from POSIX 2008 and means "run the exec'd command with as many file names as convenient" whereas the older alternative ';' (or \\; ) means "run the exec'd command once per file name".

If you have GNU find , you can use various dialects of regular expression:

find "$ROOT_DIR" -regex '.*\.\(zip\|bz2\)' -mtime +5 -delete

This uses the default regex mode; you can use some alternatives to avoid using so many backslashes. The -delete option uses the unlink() system call rather than invoking an external command; it is more efficient, therefore.