Regex subsequence matching

Question

I'm using python but code in any language will do as well for this question.

Suppose I have 2 strings.

sequence ='abcd'
string = 'axyzbdclkd'

In the above example sequence is a subsequence of string

How can I check if sequence is a subsequence of string using regex? Also check the examples here for difference in subsequence and subarray and what I mean by subsequence.

The only think I could think of is this but it's far from what I want.

import re
c = re.compile('abcd')
c.match('axyzbdclkd')

Yea but had some other work so wouldn't have been able to respond that's why closed it and asked now as I've got some time on hand :-P.. — Abhishek J, Mar 02 '17 at 12:32

score 9 · Accepted Answer · answered Mar 02 '17 at 12:27

9

Just allow arbitrary strings in between:

c = re.compile('.*a.*b.*c.*d.*')
# .* any character, zero or more times

answered Mar 02 '17 at 12:27

user2390182

72,016
6
67
89

This will only work if the `sequence` does not contain any special characters. – Willem Van Onsem Mar 02 '17 at 12:29
@WillemVanOnsem Yea true but I'm having just lowercase characters – Abhishek J Mar 02 '17 at 12:31
If the pattern grows longer, and the input is big, backtracking will most likely cause a "crash". – Wiktor Stribiżew Mar 02 '17 at 12:40
@WiktorStribiżew The pattern would be max around 20 characters and the string would be around 50-100 characters.. Would it be alright?.. – Abhishek J Mar 02 '17 at 12:42

score 3 · Answer 2 · edited Sep 26 '17 at 22:02

You can, for an arbitrary sequence construct a regex like:

import re

sequence = 'abcd'
rgx = re.compile('.*'.join(re.escape(x) for x in sequence))

which will - for 'abcd' result in a regex 'a.*b.*c.*d'. You can then use re.find(..):

the_string = 'axyzbdclkd'
if rgx.search(the_string):
    # ... the sequence is a subsequence.
    pass

By using re.escape(..) you know for sure that for instance '.' in the original sequence will be translated to '\.' and thus not match any character.

score 0 · Answer 3 · answered May 25 '18 at 23:13

I don't think the solution is as simple as @schwobaseggl claims. Let me show you another sequence from your database: ab1b2cd. By using the abcd subsequence for pattern matching you can get 2 results: ab(1b2)cd and a(b1)b(2)cd. So for testing purposes the proposed ^.*a.*b.*c.*d.*$ is ok(ish), but for parsing the ^a(.*)b(.*)cd$ will always be greedy. To get the second result you'll need to make it lazy: ^a(.*?)b(.*)cd$. So if you need this for parsing, then you should know how many variables are expected and to optimize the regex pattern you need to parse a few example strings and put the gaps with capturing groups only to the positions you really need them. An advanced version of this would inject the pattern of the actual variable instead of .*, so for example ^ab(\d\w\d)cd$ or ^a(\w\d)b(\d)cd$ in the second case.

Regex subsequence matching

3 Answers3

Linked