Implementation of Text Indexing System in Web-Based Document Search Application Using Mongodb

Abstract

The rapid growth of information technology has led to an increase in the amount of data stored in databases every day. Relational databases (SQL) that have been in use for a long time are now being developed with the emergence of NoSQL databases such as MongoDB. MongoDB stores data in BSON format and has a Text Indexes feature that is useful for speeding up text search on string content. This feature is particularly useful in searching for data in the form of texts or strings in large quantities. MongoDB's Text Indexes have a flexible schema that does not require a strict schema structure to index text data, unlike SQL databases that require columns with the appropriate data type to perform indexing. MongoDB's Text Indexes support more languages than SQL because they use an open-source text search engine called Apache Lucene. In this study, the researcher will implement Text Indexing on document data (PDF) that has been converted into text, then inserted into MongoDB before indexing. Afterward, the researcher will compare the performance of search queries between indexed and non-indexed data in MongoDB in terms of speed. The comparison results will be presented in tables and graphs to facilitate understanding. Based on the research conducted, it can be concluded that the use of the text indexing feature in MongoDB can speed up keyword or string search time. In the experiment conducted using 5000 data records, the results showed that the use of text indexing for searching 1 keyword resulted in a search speed improvement of 11705,88%, for searching 2 keywords it was 60833,33%, and for searching 3 keywords it was 44320%.