Basics of searching for data considering typos and partial results

Question

I have a large database (firebird) containing address information for every address on my country (Brazil). This will be used to locate addresses inputed by users.

My main concern is how to implement a search engine capable of resolving typos, partial results, or results that are similar to the search query.

I'm pretty clueless as to where I would begin to implement this kind of capability. What are the terminologies I should be looking for? Is there any kind of service on the internet that would provide these capabilities, given that I provide the data?

This is called address standardization. There are third-party vendors that can handle this in many countries, presumably Brazil as well. — Gordon Linoff, Mar 28 '16 at 18:45
@GordonLinoff it's not actually a problem specific to addresses per say, but to data querying. For example, if the user inputs "plista", I want the system to be able to guess he meant "paulista" — RudiBR, Mar 28 '16 at 18:52

score 3 · Accepted Answer · edited May 23 '17 at 12:31

3

You basically need Free text search capability

All the info is here Full Text Search with Firebird and Delphi

edited May 23 '17 at 12:31

Community

1
1

answered Mar 28 '16 at 19:16

elirevach

404
2
7

score 1 · Answer 2 · answered Mar 28 '16 at 19:03

You may need to know Wildcards and Regular Expressions, and Fuzzy Matching. You can google for a lot of information. You may also need to know Trie.

Those above are very complicated algorithms. But there are libraries which you can use for different languages like Python.

Basics of searching for data considering typos and partial results

2 Answers2