bash重命名不同目录中文件名的一部分

Question

This is simple and has been asked/answered many times but not when files are in different directories. I want to do this: Renaming part of a filename , but with the files in different sub-directories.

This is what I've tried:

for f in */*run?.t; do mv -i -- "$f" "${f//subset/subset5}"; done

but it tries to rename the directory too (because subset is also in the directory name) and then errors because that directory name is not found. What am I missing?

Here is the general directory structure:

Gene1_subset/
       Gene1file_subset.run1.t
       Gene1file_subset.run2.t
Gene2_subset/
       Gene2file_subset.run1.t
       Gene2file_subset.run2.t

Here is what I'd like:

Gene1_subset/
           Gene1file_subset5.run1.t
           Gene1file_subset5.run2.t
Gene2_subset/
           Gene2file_subset5.run1.t
           Gene2file_subset5.run2.t

But it attempts to rename the directories too so you get the error:

cannot move Gene1_subset/Gene1file_subset.run1 to Gene1_subset5/Gene1file_subset5.run1 : No such file or directory

Answer 1

You can check if the file is a directory and, if not, rename it.

For example, [ -d FILE ] will return true if FILE exists and is a directory.

Instead of trying to move everything from the root, cycle all directories and for each one, cycle all files.

EG

#!/bin/bash
rename() {
  for dir in *; do
    if [ -d "${dir}" ]; then
      cd "${dir}"
      for f in *run?; do
        ! [ -d "${f}" ] && mv -i -- "${f}" "${f//subset/subset5}"
      done
      cd ..
    fi
  done
}
rename || exit $?
exit 0

Answer 2

$ mkdir -p Gene1_subset Gene2_subset
$ touch Gene1_subset/Gene1file_subset.run1 Gene1_subset/Gene1file_subset.run2 Gene2_subset/Gene2file_subset.run1 Gene2_subset/Gene2file_subset.run2

I think it all comes down to finding the proper regex. Here I substitute the string subset.run for subset5.run :

$ find . -name '*.run[0-9]' -exec sh -c 'mv -v "{}" "$(echo {} | sed "s/subset\.run/subset5\.run/")"' \;
renamed './Gene2_subset/Gene2file_subset.run2' -> './Gene2_subset/Gene2file_subset5.run2'
renamed './Gene2_subset/Gene2file_subset.run1' -> './Gene2_subset/Gene2file_subset5.run1'
renamed './Gene1_subset/Gene1file_subset.run2' -> './Gene1_subset/Gene1file_subset5.run2'
renamed './Gene1_subset/Gene1file_subset.run1' -> './Gene1_subset/Gene1file_subset5.run1'

But we can for example use basename and dirname to strip the path from filename and directory and run sed only on filename:

$ find . -name '*.run[0-9]' -exec sh -c "mv -v \"{}\" \"\$(dirname \"{}\")/\$(basename \"{}\" | sed 's/subset/subset5/')\"" \;
renamed './Gene2_subset/Gene2file_subset.run2' -> './Gene2_subset/Gene2file_subset5.run2'
renamed './Gene2_subset/Gene2file_subset.run1' -> './Gene2_subset/Gene2file_subset5.run1'
renamed './Gene1_subset/Gene1file_subset.run2' -> './Gene1_subset/Gene1file_subset5.run2'
renamed './Gene1_subset/Gene1file_subset.run1' -> './Gene1_subset/Gene1file_subset5.run1'

A little escaping to learn. This will probably be more readable to you:

for f in */*run?; do f2=$(basename "$f"); mv -vi -- "$f" "$(dirname "$f")/${f2//subset/subset5}"; done
renamed 'Gene1_subset/Gene1file_subset.run1' -> 'Gene1_subset/Gene1file_subset5.run1'
renamed 'Gene1_subset/Gene1file_subset.run2' -> 'Gene1_subset/Gene1file_subset5.run2'
renamed 'Gene2_subset/Gene2file_subset.run1' -> 'Gene2_subset/Gene2file_subset5.run1'
renamed 'Gene2_subset/Gene2file_subset.run2' -> 'Gene2_subset/Gene2file_subset5.run2'

Answer 3

Something like this should do:

#!/usr/bin/env bash

shopt -s extglob globstar nullglob
for f in **/*_subset.run+([0-9]).t; do
    tmp=${f##*_}
    mv -i -- "$f" "${f%_*}_${tmp/et/et5}"
done

---

If your files strictly follow the structure, you could just change your code to:

mv -i -- "$f" "${f/file_subset/file_subset5}"