How to read a csv file with python

5
votes

I'm trying to read a csv file but it doesn't work. I can read my csv file but when I see what I read, there where white space between values.

Here is my code 

# -*- coding: iso-8859-1 -*-
import sql_db, tmpl_macros, os
import security, form, common

import csv

class windows_dialect(csv.Dialect):
    """Describe the usual properties of unix-generated CSV files."""
    delimiter = ','
    quotechar = '"'
    doublequote = 1
    skipinitialspace = 0
    lineterminator = 'n'
    quoting = csv.QUOTE_MINIMAL


def reco(d):
 cars = {210:'"', 211:'"', 213:"'", 136:'à', 143:'è', 142:'é'}
 for c in cars:
  d = d.replace(chr(c),cars[c])
 return d

def page_process(ctx):
 if ctx.req_equals('catalog_send'):
  if 'catalog_file' in ctx.locals.__dict__:
   contenu = ctx.locals.catalog_file[0].file.read()
   #contenu.encode('')
   p = csv.reader(contenu, delimiter=',')
   inserted = 0
   modified = 0
   (cr,db) = sql_db.cursor_get()
   for line in p:
    if line:
     logfile = open('/tmp/test.log', 'a')
     logfile.write(line[0])
     logfile.write('\n')
     logfile.write('-----------------------------\n')
     logfile.close()
3
pythoncsv
+1 for 'unix-generated CSV files' (showing that Excel isn't the only way) - pavium
your indentation is inconsistent. - SilentGhost
What does "whitespace between values" mean? The file has whitespace in it? The values that csv got from the file have whitespace around them? Your line terminator is wrong, BTW. It should be \n not n. - S.Lott
Possible duplicate of How do I read and write CSV files with Python? - Martin Thoma

3 Answers

7
votes

I prefer to use numpy's genfromtxt rather than the standard csv library, because it generates numpy's recarray, which are clean data structures to store data in a table-like object.

>>> from numpy import genfromtxt
>>> data = genfromtxt(csvfile, delimiter=',', dtype=None)
# data is a table-like structure (a numpy recarray) in which you can access columns and rows easily
>>> data['firstcolumn']
<content of the first column>

EDIT: This answer is quite old. While numpy.genfromtxt, nowadays most people would use pandas:

>>> import pandas as pd
>>> pd.read_csv(csvfile)

This has the advantage of creating pandas.DataFrame, which is a better structure for data analysis.

2
votes

If you have control over the data, use tab-delimited instead::

import csv
import string

writer = open('junk.txt', 'wb')
for x in range(10):
    writer.write('\t'.join(string.letters[:5]))
    writer.write('\r\n')
writer.close()
reader = csv.reader(open('junk.txt', 'r'), dialect='excel-tab')
for line in reader:
    print line

This produces expected results.

A tip for getting more useful feedback: Demonstrate your problem through self-contained and complete example code that doesn't contain extraneous and unimportant artifacts.

0
votes

You don't do anything with the dialect you've defined. Did you mean to do this:

csv.register_dialect('windows_dialect', windows_dialect)
p = csv.reader(contenu, dialect='windows_dialect')

Also not sure what the reco function is for.