Repository navigation
Allow False to be passed to filter #70027
Description
| BPO | 25840 |
|---|---|
| Nosy | @bitdancer, @leewz, @Vgr255 |
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 = <Date 2015-12-11.15:45:34.289>
created_at = <Date 2015-12-11.15:02:00.929>
labels = ['type-feature']
title = 'Allow `False` to be passed to `filter`'
updated_at = <Date 2015-12-11.19:54:59.059>
user = 'https://github.com/leewz'bugs.python.org fields:
activity = <Date 2015-12-11.19:54:59.059>
actor = 'leewz'
assignee = 'none'
closed = True
closed_date = <Date 2015-12-11.15:45:34.289>
closer = 'r.david.murray'
components = []
creation = <Date 2015-12-11.15:02:00.929>
creator = 'leewz'
dependencies = []
files = []
hgrepos = []
issue_num = 25840
keywords = []
message_count = 4.0
messages = ['256218', '256219', '256222', '256231']
nosy_count = 3.0
nosy_names = ['r.david.murray', 'leewz', 'abarry']
pr_nums = []
priority = 'normal'
resolution = 'rejected'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue25840'
versions = ['Python 3.6']
Activity
leewz commentedon Dec 11, 2015
Meaning:
I understand that it is a very minor enhancement, and with not much benefit. I just happened to think about it, and wondered why it didn't already exist. I figured it wouldn't hurt to put the idea out here.
(If anyone is interested, I was looking into ways that filter/map/itertools could "unwrap" each other at the C level to improve composition of generators, inspired by the functools.partial optimization.)
Vgr255 commentedon Dec 11, 2015
Do you mean like 'filter(None, lst)' does?
bitdancer commentedon Dec 11, 2015
I don't think it is worth adding as a special case. Itertools has it because of what itertools is (a mini-language for manipulating iterables), but the legacy filter function has no reason to grow additional special cases beyond None. (It's not even clear why it has that one special case :)
leewz commentedon Dec 11, 2015
ebarry, note that
filter(None, lst)is equivalent tofilter(bool, lst), which is the opposite offilterfalse(None, lst). (Thoughfilter(True, lst) == filter(bool, lst)would be a parallel.)