I have a directory of 170,000+ pickle files in multiple subdirectories which were originally pickled using the (protocol=0) format. This hasn't been very efficient time or space-wise.
I wrote a script to re-pickle (using cPickle, protocol=2) each file in the folder(s) but curiously, the script throws an exception while processing a particular file (file # 95,000). Initially, I thought that the pickle file is corrupted. When I try to load this exact pickle file from IPython command line, the file loads just fine.
So, I'm dumbfounded as to why this happens. Here's my script and I appreciate help:
import os
import cPickle
import numpy
import time
import re
from progressbar import ProgressBar
inpath = '/path/to/folder'
def list_files(dir):
r = []
subdirs = [x[0] for x in os.walk(dir)]
for subdir in subdirs:
files = os.walk(subdir).next()[2]
if (len(files) > 0):
for file in files:
r.append(subdir + "/" + file)
return r
infileList = list_files(inpath)
print "Total number of files found: %d" % len(infileList)
print "\n\n"
progress = ProgressBar()
outfilename = " "
print "Processing pickle files. Pls wait..."
t0 = time.time()
filecount = 0
for file in progress(infileList):
try:
arr = cPickle.load(open(file , "rb" ))
outfilename = re.sub('/initial/path/','/new/path/',file)
if not os.path.exists(os.path.dirname(outfilename)):
os.makedirs(os.path.dirname(outfilename))
with open(outfilename, "wb") as f:
cPickle.dump(arr,f,protocol=2)
filecount = filecount + 1
except Exception,e:
print "\n" + str(filecount)
print "\nError occured while processing file: " + outfilename
tx = time.time()
print "\n Time elapsed: %.2f" % (tx-t0)
continue
t1 = time.time()
total = t1-t0
print "Files repickled with protocol=2.\nRepickling execution time: %.2f sec" % total
