csv.DictReader, skipinitialspace does not ignore tabs

[andrelehmann]

BPO	35816
Nosy	@tirkarthi
Files	conf.csv

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = None
created_at = <Date 2019-01-24.10:28:34.304>
labels = ['type-bug', 'library', '3.10']
title = 'csv.DictReader, skipinitialspace does not ignore tabs'
updated_at = <Date 2021-03-25.22:59:07.428>
user = 'https://bugs.python.org/andrelehmann'

bugs.python.org fields:

activity = <Date 2021-03-25.22:59:07.428>
actor = 'iritkatriel'
assignee = 'none'
closed = False
closed_date = None
closer = None
components = ['Library (Lib)']
creation = <Date 2019-01-24.10:28:34.304>
creator = 'andre.lehmann'
dependencies = []
files = ['48075']
hgrepos = []
issue_num = 35816
keywords = []
message_count = 3.0
messages = ['334289', '334290', '334295']
nosy_count = 2.0
nosy_names = ['xtreak', 'andre.lehmann']
pr_nums = []
priority = 'normal'
resolution = None
stage = None
status = 'open'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue35816'
versions = ['Python 3.10']

[andrelehmann]

When using the csv.DictReader a dialect can be given to change the behavior of interpretation of the csv file.

The Dialect has an option "skipinitialspace" which shall ignore the whitespace after the delimiter according to the documentation (https://docs.python.org/3/library/csv.html).

Unfortunately this works only for spaces but not for tabs which are also whitespaces.

See the following code snippet applied on the attached file:

with open("conf-csv", "r") as csvfile:
    csv.register_dialect("comma_and_ws", skipinitialspace=True)
    csv_dict_reader = csv.DictReader(csvfile, dialect="comma_and_ws")
    for line in csv_dict_reader:
        print(line)

The second line shall not contain "\t" chars.

[tirkarthi]

https://bugs.python.org/issue21297#msg216907 to be related and has a patch. It refers to whitespace as only space ('U+0020') with tabs being ignored.

Current code where only space is taken into account :

cpython/Modules/_csv.c

Line 621 in fd628cf

else if (c == ' ' && dialect->skipinitialspace)

[tirkarthi]

Sorry, I overlooked the patch. The issue reported is the same in bpo-21297 but the patch was about changing whitespace to space in the doc instead of changing the behavior as I can see from the discussion.

[serhiy-storchaka]

Well, this was fixed in other way in bpo-21297/gh-65496.