# Substring search?

**URL:** <https://www.ruby-forum.com/t/substring-search/121173>\
**Category:** Ferret\
**Created:** [November 24, 2007, 10:14pm UTC](https://www.ruby-forum.com/t/substring-search/121173 "2007-11-24T22:14:39Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![dawillis](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/d/ebca7d/32.png) [@dawillis](https://www.ruby-forum.com/u/dawillis)\
**Post date:** [November 24, 2007, 10:14pm UTC](https://www.ruby-forum.com/t/substring-search/121173/1 "2007-11-24T22:14:39Z")

</div>

Is it possible to use Ferret to do substring searches efficiently? If  
not, what can I use?

Problem: 1 million+ strings need to be matched with 1 million+  
substrings. For example:

iliketotraveltohawaii  
travelmagazine

will both be matched with the substring “travel” but only the first will  
match with “hawaii”.

What I have tried:

Used Ferret to create an index with a WhiteSpaceAnalyzer by splitting  
each string into characters. travelmagazine -\> t r a v e l m a g a z i n  
e

This works, and generates the index very quickly but the search  
(PhraseQuery) is very slow. Like 200-300 ms.

I’m concerned that either Ferret is the wrong tool for this or I’m just  
taking the wrong approach to this problem.

---

<div class="post-metadata">

**Author:** ![dawillis](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/d/ebca7d/32.png) [@dawillis](https://www.ruby-forum.com/u/dawillis)\
**Post date:** [November 25, 2007, 3:37am UTC](https://www.ruby-forum.com/t/substring-search/121173/2 "2007-11-25T03:37:41Z")

</div>

To answer my own question… using n-grams instead of unigrams provided  
a huge speedup.