正则表达式 | 空字符串匹配 | Python 3.4.0

Question

My Code:

import re

#Phone Number regex
phoneRegex = re.compile(r'''(
(\d{3}|\(\d{3}\))?
(\s|-|\.)?                       # separator
(\d{3})                          # first 3 digits
(\s|-|\.)                        # separator
(\d{4})                          # last 4 digits
(\s*(ext|x|ext.)\s*(\d{2,5}))?   # extension
)''', re.VERBOSE)

phoneRegex.findall('Phone: 800.420.7240 or +1 415.863.9900 (9 a.m. to 5 p.m., M-F, PST)')

Output:

[('800.420.7240', '800', '.', '420', '.', '7240', '', '', ''), ('415.863.9900', '415', '.', '863', '.', '9900', '', '', '')]

Questions:

1. Why are empty strings included in the match?
2. The empty strings are matched from what positions of the string?
3. What are the conditions for empty strings to be matched?

PS The empty strings are not included in the match when I use the same regex on https://regex101.com/
Also, I just started learning regex a few days ago, so I'm sorry if my questions aren't good enough.

Answer 1

The ? operator means that it will return zero or one matches. In this case, you've made some capturing groups optional with ? , and python is returning a zero-length match for each of the three optional capturing groups you created.

If you remove the first two ? you'll eliminate some zero-length matches. To deal with the last one, you need to change the extension pattern. It accounts for two, again because you use a zero or one operator ( * ).

If you don't care about the internal capturing groups and just want the full match, you could filter out the zero-length matches with something like

>>> [match.group(0) for match in phoneRegex.finditer('Phone: 800.420.7240 or +1 415.863.9900 (9 a.m. to 5 p.m., M-F, PST)')]
['800.420.7240', '415.863.9900']

You could make the extension capture group match conditional on there being a match for a preceding phone number. Also, I think you may need to escape the . in the third alternative ext. . As written it matches any character, but I think what you meant was ext\. .

for reference:

• zero-length regex matches
• python re

Answer 2

Why are empty strings included in the match? Because you have used various groups in your regex. The engine captures the part of the match that you have put into a group.

The empty strings are matched from what positions of the string? From this regex: (\s*(ext|x|ext.)\s*(\d{2,5}))? It has three groups (you can count the opening parentheses). The engine does not find a match with extensions and the 3 groups that try to capture information return empty strings.

What are the conditions for empty strings to be matched? If you group your regex in a way that the engine catches an empty substring in the string matched, it will return empty strings:-)

I think you are following the exercise from "automate the boring stuff with python". In the regex in VERBOSE mode on page 178, try to look for the opening parentheses. Where is the closing parentheses? The number of groups is equivalent with the number of opening parentheses. The whole regex is group number zero.

Groups are useful if you want to extract certain parts of a matched string. If you only want to extract full phone numbers, leave the groups away.

You may try just this:

phoneRegex = re.compile(r'\d{3}[\.|-|\/]\d{3}[\.|-|\/]\d{4}')

Is this what you want to achieve?

If you want to stick with your regex in VERBOSE mode, you could also use non-capturing groups. This only captures a full match:

phoneRegex = re.compile(r'''(
(?:\d{3}|\(?:\d{3}\))?
(?:\s|-|\.)?                       # separator
(?:\d{3})                          # first 3 digits
(?:\s|-|\.)                        # separator
(?:\d{4})                          # last 4 digits
(?:\s*(?:ext|x|ext.)\s*(?:\d{2,5}))?   # extension
)''', re.VERBOSE)