比较同一文件夹下相同大小的文件。删除相同的文件，备份到制定文件夹下

比较同一文件夹下相同大小的文件。发现相同大小文件，则移动到temp_folder文件夹下。 也可以添加比较设备号，创建时间等。

find_del_same_files.py

__author__ = 'wulong'

import sys
import os
import shutil


def get_file_list(path):
    f_l = []
    
    if os.path.isdir(path):
        for tmp in os.listdir(path):
            tmp = path + os.sep + tmp
            if os.path.isdir(tmp):
                ret_list = get_file_list(tmp)
                f_l.extend(ret_list)
            else:
                f_l.append(tmp)
                
    else:
        f_l.append(path)
    return f_l


def del_same_file(file_list,bakup_dir):
    remain_list = []
    count = 1
    bakup_dir.strip()

    tmp_dir_count = 1

    tmp_dir = bakup_dir
    if (os.path.exists(tmp_dir)):
        while True:
            if (os.path.exists(tmp_dir)):
                    tmp_dir = bakup_dir + str(tmp_dir_count)
                    tmp_dir_count = tmp_dir_count + 1
            else:
                shutil.move(bakup_dir, tmp_dir)
                break

    os.makedirs(bakup_dir)

    while len(file_list) > 0:
        same_flag = False
        f_old = file_list.pop(0)
        if os.path.isfile(f_old):
            old_stat = os.stat(f_old)

            # print "old 0", f_old, ", ",old_stat.st_ino
            # print "old 1", f_old, ", ",old_stat.st_dev

            for f_remain in remain_list:
                remain_stat = os.stat(f_remain)
                if (old_stat.st_size == remain_stat.st_size) and (old_stat.st_dev == remain_stat.st_dev):
                    f_n = os.path.basename(f_old)
                    tmp = bakup_dir + os.sep + str(count) + "_" + f_n
                    while True:
                        tmp = bakup_dir + os.sep + str(count) + "_" + f_n
                        if os.path.isfile(tmp):
                            count = count + 1
                        else:
                            break

                    print "old ", old_stat.st_size, ", ",f_old
                    print "remain ",remain_stat.st_size, ", ",f_remain

                    shutil.move(f_old, tmp)
                    same_flag = True
                    count = count + 1
                    break
            if (not same_flag):
                remain_list.append(f_old)


old_path=""

if len(sys.argv) == 2:
    old_path = sys.argv[1]

old_path.strip()


if os.path.isdir(old_path):
    os.chdir(old_path)
    old_path = os.getcwd()
    old_file_list = get_file_list(old_path)
    del_same_file(old_file_list, old_path + os.sep + "temp_folder")
else:
    print "param is not directory ",old_path