python - Splitting a CSV file into equal parts? -


i have large csv file split number equal number of cpu cores in system. want use multiprocess have cores work on file together. however, having trouble splitting file parts. i've looked on google , found sample code appears want. here have far:

def split(infilename, num_cpus=multiprocessing.cpu_count()):     read_buffer = 2**13     total_file_size = os.path.getsize(infilename)     print total_file_size     files = list()     open(infilename, 'rb') infile:         in xrange(num_cpus):             files.append(tempfile.temporaryfile())             this_file_size = 0             while this_file_size < 1.0 * total_file_size / num_cpus:                 files[-1].write(infile.read(read_buffer))                 this_file_size += read_buffer         files[-1].write(infile.readline()) # possible remainder         files[-1].seek(0, 0)     return files  files = split("sample_simple.csv") print len(files)  ifile in files:     reader = csv.reader(ifile)     row in reader:         print row

the 2 prints show correct file size , split 4 pieces (my system has 4 cpu cores).

however, last section of code prints rows in each of pieces gives error:

for row in reader: _csv.error: line contains null byte

i tried printing rows without running split function , prints values correctly. suspect split function has added null bytes resulting 4 file pieces i'm not sure why.

does know if correct , fast method split file? want resulting pieces can read csv.reader.

as said in comment, csv files need split on row (or line) boundaries. code doesn't , potentially breaks them somewhere in middle of 1 — suspect cause of _csv.error.

the following avoids doing processing input file series of lines. i've tested , seems work standalone in sense divided sample file approximately equally size chunks because it's unlikely whole number of rows fit chunk.

update

this substantially faster version of code posted. improvement because uses temp file's own tell() method determine changing length of file it's being written instead of calling os.path.getsize(), eliminated need flush() file , call os.fsync() on after each row written.

import csv import multiprocessing import os import tempfile  def split(infilename, num_chunks=multiprocessing.cpu_count()):     read_buffer = 2**13     in_file_size = os.path.getsize(infilename)     print 'in_file_size:', in_file_size     chunk_size = in_file_size // num_chunks     print 'target chunk_size:', chunk_size     files = []     open(infilename, 'rb', read_buffer) infile:         _ in xrange(num_chunks):             temp_file = tempfile.temporaryfile()             while temp_file.tell() < chunk_size:                 try:                     temp_file.write(infile.next())                 except stopiteration:  # end of infile                     break             temp_file.seek(0)  # rewind             files.append(temp_file)     return files  files = split("sample_simple.csv", num_chunks=4) print 'number of files created: {}'.format(len(files))  i, ifile in enumerate(files, start=1):     print 'size of temp file {}: {}'.format(i, os.path.getsize(ifile.name))     print 'contents of file {}:'.format(i)     reader = csv.reader(ifile)     row in reader:         print row     print ''

Comments

Post a Comment

Popular posts from this blog

powershell Start-Process exit code -1073741502 when used with Credential from a windows service environment -

i'm running strange behavior powershell start-process call. here call: $process = start-process ` "c:\somepath\mybinary.exe" ` -passthru ` -credential $defaultcredential ` -wait ` -workingdirectory "c:\somepath" ` -loaduserprofile if ($process.exitcode -ne 0) { #do } this call return exit code - 1073741502 . after quick search, exit code seems related generic error when program not load required dll (aka. status_dll_init_failed ). when run without -credential $credential program runs correctly. in order isolate problem, manually launched some.exe in prompt target credential , runs smoothly. so problem seems come way start-process cmdlet launch process. i found potential solutions problem tried apply no luck : link , link . would have idea of what's going on here ? edit 1: run proc mon monitoring program activities when launched directly or via powershell script. problem seems occur when loading kernelbase
Read more

twig - Using Twigbridge in a Laravel 5.1 Package -

i'm trying create laravel 5.1 package uses twigbridge instead of .php views. can normal .php views display fine, when try make use of .twig templates i'm met error. possible use twigbridge templates when creating views laravel package or pretty stuck blade , php ? edit 1: loading twig template without extending works. it's when start using {% extends 'file' %} error comes out. i error error loading /virtual/laravel/packages/username/mypackage/src/views/sample.twig: template "templates.master" not defined () in "/virtual/laravel/packages/username/mypackage/src/views/sample.twig" @ line 1. my folders username |_ mypackage |_ src |_ controllers - samplecontroller.php |_ views |_ templates - master.twig - sample.twig twig template sample.twig {% extends 'templates.master' %} {% block content %} {{ parent()
Read more

c# - LINQ join Entities from HashSet's, Join vs Dictionary vs HashSet performance -

i have hashset's each stores t, have written test application compares different relation algorithms can think of, i'm not pleased results getting. does there exist more efficient ways of achieving object relations, once testing? using system; using system.collections.generic; using system.linq; using system.text; using system.threading.tasks; using system.diagnostics; namespace linqtests { class program { static void main(string[] args) { hashset<user> usertable = new hashset<user>(); hashset<userproperty> userpropertytable = new hashset<userproperty>(); #region lets create dummy data. console.writeline("please wait while creating dummy data, can take while..."); console.writeline(""); int rows = 1000000; for(int x = 0; x < rows; x++) { random rnd = new random(); // ad
Read more