使用 Python 使用正则表达式查找重叠序列

Question

I'm trying to extract numbers and both previous and following characters (excluding digits and whitespaces) of a string. The expected return of the function is a list of tuples, with each tuple having the shape:

(previous_sequence, number, next_sequence)

For example:

string = '200gr T34S'
my_func(string)
>>[('', '200', 'gr'), ('T', '34', 'S')]

My first iteration was to use:

def my_func(string):
    res_obj = re.findall(r'([^\d\s]+)?(\d+)([^\d\s]+)?', string)

But this function doesn't do what I expect when I pass a string like '2AB3' I would like to output [('','2','AB'), ('AB','3','')] and instead, it is showing [('','2','AB'), ('','3','')] , because 'AB' is part of the previous output.

How could I fix this?

Answer 1

Instead of modifier + and ? you can simply use * :

>>> re.findall(r'([^\d\s]*)(\d+)([^\d\s]*)',string)
[('', '200', 'gr'), ('T', '34', 'S')]

But if you mean to match the overlapped strings you can use a positive look ahead to fine all the overlapped matches :

>>> re.findall(r'(?=([^\d\s]*)(\d+)([^\d\s]*))','2AB3')
[('', '2', 'AB'), ('AB', '3', ''), ('B', '3', ''), ('', '3', '')]

Answer 2

Since there is no overlapping numbers, a single trailing
assertion should be all you need.

Something like ([^\\d\\s]+)?(\\d+)(?=([^\\d\\s]+)?)

This ([^\\d\\s]*)(\\d+)(?=([^\\d\\s]*)) if you care about
the difference between NULL and the empty string.

Answer 3

Another way can be using regex and functions!

import re

#'200gr T34S'  '2AB3'


def s(x):
    tmp=[]
    d = re.split(r'\s+|(\d+)',x)
    d = ['' if v is None else v for v in d] #remove None
    t_ = [i for i in d if len(i)>0]
    digits = ['0', '1', '2', '3', '4', '5', '6', '7', '8', '9']
    nms = [i for i in t_ if i[0] in digits]
    for i in nms:       
        if d.index(i)==0:
            tmp.append(('',i,d[d.index(i)+1]))
        elif d.index(i)==len(d):
            tmp.append((d[d.index(i)-1],i,''))
        else:
            tmp.append((d[d.index(i)-1],i,d[d.index(i)+1]))
    return tmp

print s('2AB3')

Prints-

[('', '2', 'AB'), ('AB', '3', '')]