Small Search Engine Algorithm for Document Word Search

Question

I have to design and implement an algorithm for my university project that searches a given set of documents based on the keywords/query given. Assume that each document contain few sentences and these documents can be stored in a suitable data structure. When a query is made I have to display the documents that contain the keywords. A query can contain simple logical operators such as "AND" and "OR".

For example assume that there are 3 documents named Doc1, Doc2, Doc3 with this content:

• Doc1: This is my University.
• Doc2: My University is situated at Delhi.
• Doc3: I like My University.

Here are the answers to some queries:

1. "University": Doc1, Doc2, Doc3
2. "my AND University": Doc1, Doc2
3. "like OR Delhi": Doc2, Doc3

Currently what I have developed reads each file and put its contents into separate binary trees, and I've developed a function for searching one word from the binary trees. How can I extend my search algorithm for search with logical operators?

• paparazzo: Your are not looking for third party tool? You need to build this from the ground up? Lucene is Open Source.
• SHdinesh Madushanka: yeah. I don't want to use any 3rd party tool and request to provide any possible approach to implement this.
• doppelgreener: If I understand correctly, the answers to some queries are theoretical expected results, not actual results you're already getting, is that right?
• jscs: Why does query 2 "my AND University" not return all three documents? Or, if this is case-sensitive, why does it return Doc2?
• Asaf: if you know to find one word then break the string into tokens and ineach step check two: the word + operator. if or and found word true, if "and" and not found... false else move to the next two query words.

Answer 1

Hash the words to a key value pair of (word, set of documents). When you do the search, insert the sets found into the hashtable. Then do a union on the sets for OR and intersections for AND

• SHdinesh Madushanka: thanks for your answer and can you provide an example?

Answer 2

You could go with .

Dictionary<string, HashSet<Int>> se
 word docID

But I thought you had to build from scratch

I don't know java
It may be called a HashTable in java

var docs = se["my"].Interset(se["university"]);
var docs = se["my"].Union(se["Delhi"]);