Python: Use UnicodeWriter to write Unicode to CSV
The Python documentation has the following code examples for writing unicode to a csv file. I think it mentions there that this is because the csv module can’t handle the methods of unicode strings.
class UnicodeWriter:
"""
A CSV writer which will write rows to CSV file "f",
which is encoded in the given encoding.
"""
def __init__(self, f, dialect=csv.excel, encoding="utf-8", **kwds):
# Redirect output to a queue
self.queue = cStringIO.StringIO()
self.writer = csv.writer(self.queue, dialect=dialect, **kwds)
self.stream = f
self.encoder = codecs.getincrementalencoder(encoding)()
def writerow(self, row):
self.writer.writerow([s.encode("utf-8") for s in row])
# Fetch UTF-8 output from the queue ...
data = self.queue.getvalue()
data = data.decode("utf-8")
# ... and reencode it into the target encoding
data = self.encoder.encode(data)
# write to the target stream
self.stream.write(data)
# empty queue
self.queue.truncate(0)
def writerows(self, rows):
for row in rows:
self.writerow(row)
I’m writing multiple files, and for simplicity I’ve just put a piece of code to demonstrate how I can use the above class in my code:
def write(self):
"""
Outputs the dataset to a csv.
"""
f = codecs.open(self.filename, 'a')
writer = UnicodeWriter(f)
#with open(self.filename, 'a', encoding='utf-8') as f:
if self.headers and not self.written:
writer.writerow(self.headers)
self.written = True
for record in self.records[self.last_written:]:
print record
writer.writerow(record)
self.last_written = len(self.records)
f.close()
This is a method inside a coll dataset that prepares the dataset before writing csv, previously I used writer = csv.writer(f) but
due to codec error I changed my code to use the ‘UnicodeWriter class.’
But my problem is that when I open the csv file, I get the following information:
some_header
B,r,ë,k,ò,w,n,i,k,_,b,s
B,r,ë,k,ò,w,n,i,k,_,c,s
B,r,ë,k,ò,w,n,i,k,_,c,s,b
B,r,ë,k,ò,w,n,i,k,_,d,e
B,r,ë,k,ò,w,n,i,k,_,d,e,-,1
B,r,ë,k,ò,w,n,i,k,_,d,e,-,2
B,r,ë,k,ò,w,n,i,k,_,d,e,-,3
B,r,ë,k,ò,w,n,i,k,_,d,e,-,4
B,r,ë,k,ò,w,n,i,k,_,d,e,-,5
B,r,ë,k,ò,w,n,i,k,_,d,e,-,M
B,r,ë,k,ò,w,n,i,k,_,e,n
B,r,ë,k,ò,w,n,i,k,_,e,n,-,1
B,r,ë,k,ò,w,n,i,k,_,e,n,-,2
The lines should actually be something like Brëkòwnik_de-1
, I’m not really going to see what’s going on.
To give a basic understanding of how the data is generated, I’ll add the following lines:
title = unicode(row_page_title['page_title'], 'utf-8')
Solution
This symptom points to something like entering a string into a function/method that requires a list or tuple.
The writerows
method requires a list of lists, and writerows
require a list (or tuple) containing field values. Since you provide it with a string, and when you iterate over it, the string can simulate a list of characters, you get a CSV with one character for each column.
If your CSV has only one column, you should use writer.writerow([data]) instead of writer.writerow(data
).
If you only have one column, some people may question if you really need a csv module, but a csv module will handle things like records with interesting content (CR/LF and others), so yes, that’s a good idea.