generate regular expressions from a list of strings

[bhargavah]

I am reading something from regex format, expands it and writing it out. This list can become huge while writing it out.

While writing it out, I do not have the original regex data. So, I will have to create regexes out of the strings which I have.

A couple of cases while reading and writing:

Say, read regular expression is:

abc/*
Since 'abc' can have only 'A', 'B', 'C', 'D'(Have this list with me), Above would be translated to list of strings as

"abc/A", "abc/B", "abc/C", "abc/D" -- 1
Say, another read regular expression is:

def/*/A
Since 'def' can have only 'x', 'y', 'x'(Have this list with me), Above would be translated to list of strings as

"def/x/A", "def/x/A", "def/x/A" -- 2
I have already said that I do not have original regular expressions now. All I have is list of strings. I will have to create regexes out of statements number 1 and 2.

From number 1, I should get

abc/*
From number 2, I should get

def/*/A
which are the original.

Question: Which data structure would be efficient to solve this problem. I have thought of using tries & Aho–Corasick algorithm but could not get a clear solution at top of my head till now.

I would be happy to expand the question more in case it is not clear. Please consider that * will not match /, //, or anything except characters.

[tpb]

Give more examples of input/output.

[dhayden]

It looks like you're trying to output the set of strings that will match the expression. That's impossible since you can easily form an expression that matches zero or more characters.

[tpb]

dhayden wrote:

It looks like you're trying to output the set of strings that will match the expression

I think he's doing the opposite. Given a set of strings, determine the single "regular expression" that describes them. The regexes are really just expressions using the shell/command prompt * character, though.