Whats the difference between 'rb' and 'rU' in the open() function for csv

Question

I was wondering what the difference between:

with codecs.open('xxxx.csv', 'rU') as h:

and

with codecs.open('xxxx.csv', 'rb') as h:

I think I remember someone saying you should use 'rb' and not 'rU' when I started my project with reading .csv files, but I cant seem to find it again.

Anybody who would like to explain this? Thanks

Answer 1

As the documentation states, U (Universal Newlines) mode is deprecated; you should not use it anymore. Instead use the newline= keyword argument.

The csv documentation states that it prefers that parameter to be '', so open won't do any interpretation of newlines and leave that up to the csv module.

Since you likely want to decode your CSV into text (not bytes), it makes no real sense to open them in b (binary) mode.

Bottomline: the usual way to parse CSV files is:

with open('eggs.csv', newline='') as csvfile:
    spamreader = csv.reader(csvfile)
    for row in spamreader:
        ...

This means you're using the implicit open mode rt, for reading in text mode. Unless you have very special needs, this is probably what you want. And the above sample code is taken straight from the documentation.