首页猿问如何使用正则表达式从文本中解析缺少...

如何使用正则表达式从文本中解析缺少子字符串的子字符串

Python

萧十郎 2022-07-12 09:36:11

我想从字符串中解析出遵循特定格式的子字符串调查回复的格式为<survey_name>_<category_name>_<question_type>.<response_type>例如：输入字符串y_survey_category1_1st.nox_survey_category2_2ndsurvey_z_category_3_3rd.yes_more_7x_survey_category_4_4th.excludedsurvey_z_category5.yes_more_7survey_z_category_6.yes_more_7这是我到目前为止所拥有的。它适用于大多数情况，除了 question_type 是可选的（例如：上面的 5 和 6 个输入）。以下是每个子部分的限制 1. survey_name can only be one of the 3 values 2. category_name will always be present and can have underscores 3. question_type may be present and may have underscore in it 4. response_type may be present and may have underscore in it 5. Either question_type or response_type or both will always be present(x_survey|y_survey|survey_z)_([\w_]+)_(1st|2nd|3rd|4th)[.]?(.*)https://regex101.com/r/bGc0gM/1有关如何修改正则表达式以使其适用于所有情况的任何帮助？

查看完整描述

2 回答

GCT1015

TA贡献1827条经验获得超4个赞

很难找到使用的正则表达式，question_type即使它是可选的，我强制类别字母/下划线并以数字结尾

category : [a-z_]+\d
all : (x_survey|y_survey|survey_z)_([a-z_]+\d)(?>_(1st|2nd|3rd|4th))?(?>\.(.*))?

Regex demo

反对回复 2022-07-12

慕森卡

TA贡献1806条经验获得超8个赞

带有一段代码的简短版本显示了一个可以工作的正则表达式。该模式添加了额外的空格，因此您需要设置re.X标志。也设置re.I为忽略大小写。

# Capture:

# <survey_name>_<category_name>_<question_type>.<response_type>

# (0) (1) (2) or (4) (3) or (6)

pat = r"""^(x_survey|y_survey|survey_z) # <sn> (0)

_ # _

([^.]+ # <cn> (1)

(?: # One of

_(1st|2nd|3rd|4th) [.]([\w]+)$ | # qt (2) & rt (3)

_(1st|2nd|3rd|4th) $ | # qt (4)

(?<!_(1st|2nd|3rd|4th)) [.]([\w]+)$ # rt (6)

)

"""

matcher = re.compile(pat, re.I | re.X)

对于包含测试用例的解决方案的两个变体的较长版本：

"""

Format: <survey_name>_<category_name>_<question_type>.<response_type>

1. survey_name can only be one of the 3 values

2. category_name will always be present and can have underscores

3. question_type may be present and may have underscore in it

4. response_type may be present and may have underscore in it

5. Either question_type or response_type or both will always be present

A) <survery_name> always there

easy to find, one of three: (x_survey|y_survey|survey_z)

B) <category_name> always there

has 0 or more internal underscores

C) <question_type> optional

one ore more internal underscores

ends before a dot or at end of line

one of 4 values: (1st|2nd|3rd|4th)

D) <response_type> optional

starts before a .

ends at end of line

Both category_name and question_type can have zero or more internal

underscores. This results in an ambiguity, since we have no way of knowing

when category_name ends and question_type starts.

Assume that question_type is one of the 4 values (1st|2nd|3rd|4th). this

results in 3 valid cases and one that should not match:

Format: <survey_name>_<category_name>_<question_type>.<response_type>

0) Both question_type and response_type present

-->

1) Only question_type and no response_type present

(x_survey|y_survey|survey_z)_<category_name>_(1st|2nd|3rd|4th)

-->

p2 = r"^(x_survey|y_survey|survey_z)_([^.]+)_(1st|2nd|3rd|4th)$"

2) No question_type and only response_type present

(x_survey|y_survey|survey_z)_<category_name>.<response_type>

-->

3) Neither question_type nor response_type present

(x_survey|y_survey|survey_z)_<category_name>

Neither of p1, p2 nor p3 will match.

Since the patterns are mutually exclusive we can try them one after the other.

We could also combine them into one pattern.

We can combine the three patterns in one large pattern or we can try them one

after the other.

"""

from collections import namedtuple

import re

Response = namedtuple('Response', ['survey_name',

'category_name',

'question_type',

'response_type'])

cases = ["survey_z__CATEGORY",

"y_survey_category1_1st.no",

"x_survey_category2_2nd",

"survey_z_category_3_3rd.yes_more_7",

"x_survey_category_4_4th.excluded",

"X_SURVEY_CATEGORY_4_4TH.excluded",

"survey_z_category5.yes_more_7",

"survey_z_category_6.yes_more_7",

"survey_z_category_7._yes_more_77",

"survey_z_category_8_._yes_more_88",

"survey_z_category_8888__foo._yes_more_77_",

"survey_z_category_22_22_1st_2nd._yes_more_77_",

"survey_z__CATEGORY_2222_1ST__2ND._yes_odd_77_",

"survey_z__CATEGORY_3333_1ST__2ND",

"survey_z__CATEGORY_foo_1ST__2ND._yes_odd_77_",

]

def parse_survey_response_1(line):

"""Parse a line with a survey response seting optional values not

present to None. Return a Response or None when no match.

Use a list of mutually exclusive patterns for line format:

<survey_name>_<category_name>_<question_type>.<response_type>

"""

# Format <sn>_<cn>_(1st|2nd|3rd|4th).<rt>

# Format <sn>_<cn>_(1st|2nd|3rd|4th)

# Format: <sn>_<cn>.<rt>

prfx = r"^(x_survey|y_survey|survey_z)_([^.]+)"

regexs = [

prfx + r"_(1st|2nd|3rd|4th)[.]([\w]+)$", # 4 captures

prfx + r"_(1st|2nd|3rd|4th)($)", # 3+1 captures

prfx + r"(?<!_(1st|2nd|3rd|4th))[.]([\w]+)$", # 4 captures

]

matchers = [re.compile(r, re.I | re.X) for r in regexs]

for m in matchers:

parsed_line = m.search(line)

if parsed_line:

map_empty2none = (g if g else None for g in parsed_line.groups())

return Response._make(map_empty2none)

return None

def parse_survey_response_2(line):

"""Parse a line with a survey response seting optional values not

present to None. Return a Response or None when no match.

Use a one large pattern for line format:

<survey_name>_<category_name>_<question_type>.<response_type>

"""

# Capture:

# <survey_name>_<category_name>_<question_type>.<response_type>

# (0) (1) (2) or (4) (3) or (6)

pat = r"""^(x_survey|y_survey|survey_z) # <sn> (0)

_ # _

([^.]+) # <cn> (1)

(?: # One of

_(1st|2nd|3rd|4th) [.]([\w]+)$ | # qt (2) & rt (3)

_(1st|2nd|3rd|4th) $ | # qt (4)

(?<!_(1st|2nd|3rd|4th)) [.]([\w]+)$ # rt (6)

)

"""

matcher = re.compile(pat, re.I | re.X)

parsed_line = matcher.search(line)

if parsed_line:

pg = list(parsed_line.groups())

pg[2] = pg[2] if pg[2] else pg[4] # capture 2 or 4

pg[3] = pg[3] if pg[3] else pg[6] # capture 3 or 6

return Response._make(pg[:4])

return None

def unparse_survey(response):

if response.response_type:

head = '_'.join(e for e in response[:-1] if e)

unparsed = '.'.join([head, response.response_type])

else:

unparsed = '_'.join(e for e in response if e)

return unparsed

for c in cases:

p1 = parse_survey_response_1(c)

p2 = parse_survey_response_2(c)

print(c)

print(p1)

print(p2)

print(20*'=')

if p1 or p2:

assert(c == unparse_survey(p1))

assert(c == unparse_survey(p2))

运行给出：

run reex02.py

survey_z__CATEGORY

None

====================

y_survey_category1_1st.no

Response(survey_name='y_survey', category_name='category1', question_type='1st', response_type='no')

====================

x_survey_category2_2nd

Response(survey_name='x_survey', category_name='category2', question_type='2nd', response_type=None)

====================

survey_z_category_3_3rd.yes_more_7

Response(survey_name='survey_z', category_name='category_3', question_type='3rd', response_type='yes_more_7')

====================

x_survey_category_4_4th.excluded

Response(survey_name='x_survey', category_name='category_4', question_type='4th', response_type='excluded')

====================

X_SURVEY_CATEGORY_4_4TH.excluded

Response(survey_name='X_SURVEY', category_name='CATEGORY_4', question_type='4TH', response_type='excluded')

====================

survey_z_category5.yes_more_7

Response(survey_name='survey_z', category_name='category5', question_type=None, response_type='yes_more_7')

====================

survey_z_category_6.yes_more_7

Response(survey_name='survey_z', category_name='category_6', question_type=None, response_type='yes_more_7')

====================

survey_z_category_7._yes_more_77

Response(survey_name='survey_z', category_name='category_7', question_type=None, response_type='_yes_more_77')

====================

survey_z_category_8_._yes_more_88

Response(survey_name='survey_z', category_name='category_8_', question_type=None, response_type='_yes_more_88')

====================

survey_z_category_8888__foo._yes_more_77_

Response(survey_name='survey_z', category_name='category_8888__foo', question_type=None, response_type='_yes_more_77_')

====================

survey_z_category_22_22_1st_2nd._yes_more_77_

Response(survey_name='survey_z', category_name='category_22_22_1st', question_type='2nd', response_type='_yes_more_77_')

====================

survey_z__CATEGORY_2222_1ST__2ND._yes_odd_77_

Response(survey_name='survey_z', category_name='_CATEGORY_2222_1ST_', question_type='2ND', response_type='_yes_odd_77_')

====================

survey_z__CATEGORY_3333_1ST__2ND

Response(survey_name='survey_z', category_name='_CATEGORY_3333_1ST_', question_type='2ND', response_type=None)

====================

survey_z__CATEGORY_foo_1ST__2ND._yes_odd_77_

Response(survey_name='survey_z', category_name='_CATEGORY_foo_1ST_', question_type='2ND', response_type='_yes_odd_77_')

====================

反对回复 2022-07-12

2 回答
0 关注
94 浏览

关注

添加回答

0/150

提交

取消

热搜

最近搜索清空

如何使用正则表达式从文本中解析缺少子字符串的子字符串

如何使用正则表达式从文本中解析缺少子字符串的子字符串

2 回答

添加回答