Question

If I have string A and I have many other strings and I want to see if any of those other strings are in A.

What algorithm could do this in as few iterations possible?

Ex:

‘Hello, my name is Bob.’

And I want to see if ‘name is b’ is contained, which it is starting at [11].

I’m not looking to use a regular expression library.

Thanks

Answer 1

The most efficient algorithm for this is the Aho-Corasick algorithm, which given a string of length n and set of strings of total length m can find all matches in time O(n + m + z), where z is the total number of matches reported. It’s based on finite automata and is a generalization of the KMP string matching algorithm.

One cool aspect of this algorithm is that if you have a fixed set of keywords and a bunch of text strings you want to search, then the algorithm can be sped up by doing O(m) preprocessing to build up the matcher. You can then find all matches in a string of length n in time O(n + z).

If, on the other hand, you have a fixed string and then want to match a varying set of pattern strings against it, consider looking into suffix trees, which give the same runtime guarantees but are faster if the text is fixed.

Hope this helps!