Jason/synology_link_manipultor
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

ready to delete duplicated files

Browse Source
This commit is contained in:
Jason Zhu 2021-08-15 19:33:40 +10:00
parent 5ed4d99f43
commit 65d20e9701
2 changed files with 11 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
deal_duplicate_files.py
View File

@ -1,4 +1,5 @@
import pathlib import pathlib
import os
import argparse import argparse
from numpy import source from numpy import source
import pandas as pd import pandas as pd
@ -15,6 +16,11 @@ if __name__ == "__main__":
type=str, type=str,
help='csv file to the directory that be compared') help='csv file to the directory that be compared')
parser.add_argument('--remove',
type=str,
help='remove duplicated file or not',
choices=['True','False'])
args = parser.parse_args() args = parser.parse_args()
csv_source = args.source csv_source = args.source
csv_target = args.target csv_target = args.target
@ -41,3 +47,6 @@ if __name__ == "__main__":
for row in df_filtered.Path: for row in df_filtered.Path:
p = pathlib.Path(row) p = pathlib.Path(row)
print(str(p) + ": " + str(p.exists())) print(str(p) + ": " + str(p.exists()))
if args.remove == "True" and p.exists() == True:
os.remove(p)
print(str(p) + " Removed")

2
mv_unique_files.py
View File

@ -101,7 +101,7 @@ if __name__ == "__main__":
if args.csv_path: if args.csv_path:
csv_path = args.csv_path csv_path = args.csv_path
df_csv.to_csv(csv_path) df_csv.to_csv(csv_path,index_label='Index')
else: else:
for x in paths: for x in paths:
print(str(x)) print(str(x))