I am probably going to feel very dumb when someone spots what I'm doing wrong here, but I am finding myself unable to defeat what looks like it should be a simple error.
I am writing some data to a CSV with Python. One of the things I want to write is a list
of integers. I join
the list into a string before writing it to the file:
with open('publishers.csv', 'wb') as f:
writer = csv.writer(f, quoting=csv.QUOTE_NONE, delimiter='|', quotechar='')
for item in big_list_of_objects:
description = item.description
number_list = item.number_list
formatted_numbers = "-".join(number_list)
writer.writerow([
description,
formatted_numbers
])
number_list
may have anywhere from zero to a whole bunch of numbers in it. If it's an empty list, the join
just sets formatted_numbers
equal to a blank string. If it's not an empty list, I get a string made of up integers connected by hyphens.
number_list = [1,2,34,12]
formatted_numbers = '1-2-34-12'
number_list = []
formatted_numbers = ''
That's the idea, anyway. In reality, what happens is the first five rows write successfully then I get:
File "<console>", line 1, in <module>
File "/path/path/path.py", line 500, in offending_function
formatted_numbers
Error: need to escape, but no escapechar set
Now in this particular situation, the first five rows that write successfully have an empty number_list
. The row that consistently crashes also has an empty number_list
. There is nothing weird about the value being written immediately before or after number_list
on this row. And there is nothing weird about the formatted_numbers
being written when this error crops up - I tossed in a print
statement to debug, and it's just an empty string like the five before it.
Can anyone help me figure out where I might be going wrong here?
Edit: I have added these print statements:
with open('publishers.csv', 'wb') as f:
writer = csv.writer(f, quoting=csv.QUOTE_NONE, delimiter='|', quotechar='')
for item in big_list_of_objects:
description = item.description
print "Description for %r is %r" % (item,description)
number_list = item.number_list
print "Now formatting %r for %r" % (number_list,item)
formatted_numbers = "-".join(number_list)
print repr(formatted_numbers)
writer.writerow([
description,
formatted_numbers
])
The result:
Description for 'p89' is u''
Now formatting '' for 'p89'
''
Description for 'p88' is u''
Now formatting '' for 'p88'
''
Description for 'p83' is u''
Now formatting '' for 'p83'
''
Description for 'p82' is u'in-tr-t91411'
Now formatting '' for 'p82'
''
Description for 'p81' is u''
Now formatting '' for 'p81'
''
Traceback (most recent call last):
File "<console>", line 1, in <module>
File "/path/path/path.py", line 501, in offending_function
formatted_numbers
Error: need to escape, but no escapechar set
p81
is not written to the CSV - this is where the crash occurs. However, as you can see, print repr(formatted_numbers)
reveals it to be a blank string identical to those before it. There is no description
for item p81
(just a blank string), but there is a description
for the item preceding it.
The issue is most probably occuring because your description
has |
in it, which is the delimiter for your csv as well. Hence, csv is trying to escape it, but cannot since no csv.escapechar
s are set. Example to show same issue in my computer -
>>> description = 'asda|sd'
>>> formatted_numbers = ''
>>> with open('a.csv','w') as f:
... writer = csv.writer(f, quoting=csv.QUOTE_NONE, delimiter='|', quotechar='')
... writer.writerow([
... description,
... formatted_numbers
... ])
...
Traceback (most recent call last):
File "<stdin>", line 5, in <module>
_csv.Error: need to escape, but no escapechar set
One fix would be to provide an escapechar so that it can be escaped. Example -
writer = csv.writer(f, quoting=csv.QUOTE_NONE, delimiter='|', quotechar='',escapechar='\\') #Or any other appropriate escapechar
Or another fix would be to remove the |
in the description before trying to write it, if you do not really need it in the description field -
description = description.replace('|','')
Or you can quote all the fields , by using csv.QUOTE_ALL
instead of csv.QUOTE_NONE
as provide a valid quotechar
.