如果您知道文件名和它所在的目录路径，那么无论如何都没有合理的理由在这里使用find。

path='/'
for dir in path to deepest possible dir containing the file; do
    path=$path/$dir
    test -e "$path/mysql_db_name" && break
done
test -e "$path/mysql_db_name" || { echo "$0: mysql_db_name not found" >&2; exit 2;}
: proceed to use "$path/mysql_db_name"

字符串
这将继续查找具有该名称的文件的最高目录。应该不难看出如何从最深的目录开始;这里是一种方法。

path=/path/to/deepest/possible/dir/containing/the/file
while [ "$path" ]; do
    test -e "$path/mysql_db_name" && break
    path=${path%/*}
done
test -e "$path/mysql_db_name" || { echo "$0: mysql_db_name not found" >&2; exit 2;}
: proceed to use "$path/mysql_db_name"

型

假设目录树不是无限深的，我们可以递归调用的shell助手函数将是有用的，它在zsh & bash中应该是一样的：

find_files_in_dir_branch() {
  # if there is a regular file in this dir, print its full path
  [ -f "$1" ] && readlink -f "$PWD/$1"

  # if we have reached the root dir, finish with an error
  if [ "$PWD" = / ] ; then 
    false
  # otherwise, run same function in a subshell, in parent dir
  else 
    (cd .. ; find_files_in_dir_branch "$1")
  fi 
}

find_farthest_file_named() {
  local candidates="$(find_files_in_dir_branch $1)"
  # are there any canditates at all? then print the last/farthest one
  [ -n "$candidates" ] && tail -1 <<< "$candidates"
}

# invoke as find_farthest_file_named [something]

字符串
把它变成一个单行代码有点长--无论如何把它放在一个函数中是实用的，因为我们可以很容易地看到当没有找到任何东西时它们是否以代码1退出。
示例用法：

/# mkdir -p /h/a/y/stack ### test dir structure
/# touch /needle /h/a/needle /h/a/y/stack/needle ### 3 needles in haystack
/# cd /h/a/y/stack/
/h/a/y/stack# find_farthest_file_named needle ; echo $? ### find root one and succeed
/needle
0
/h/a/y/stack# rm /needle 
/h/a/y/stack# find_farthest_file_named needle ; echo $? ### find middle one and succeed
/h/a/needle
0
/h/a/y/stack# rm /h/a/needle
/h/a/y/stack# find_farthest_file_named needle ; echo $? ### find last one and succeed
/h/a/y/stack/needle
0
/h/a/y/stack# rm needle 
/h/a/y/stack# find_farthest_file_named needle ; echo $? ### find none and fail
/h/a/y/stack# echo $?
1

型