Make Database Queries Without the Database

By Brian Goetz
2005-07-14


My kingdom for a data model
At this point, the lack of a formal data model, which describes the data gathered and against which you could more easily express selection and aggregation queries, was starting to look like a liability. Perhaps laziness was not as efficient as first hoped. But, while this application had no formal data model, maybe we could "borrow one" by streaming the data into an in-memory database and querying against that. Two possibilities immediately sprang to mind; HSQLDB, the open-source in-memory SQL database, and XQuery. I didn't need the persistence that a database offered, but I did want the query language.

HSQLDB is an embeddable database engine written in the Java™ language. It has table types for both in-memory and disk-based tables, and was designed for being embedded entirely within an application, eliminating the administrative overhead associated with most real databases. To load the data into HSQLDB, it would only be necessary to write a Visitor that traversed the in-memory data structure and generated the appropriate INSERT statements for each entity to be stored. Then you could execute SQL queries against the in-memory database tables to do the reports and throw away the "database" when finished.

Oops, forgot how annoying relational databases are
The HSQLDB approach was workable, but it quickly became clear that I would have to pay the object-relational impedance mismatch penalty not once, but twice -- once when converting the tree-structured database into a relational data model, and again when trying to turn the results of flat relational queries into structured XML or HTML result sets. And post-processing the JDBC ResultSet into a DOM representation of an XML or HTML document was still nontrivial and would require some customized coding for each report. So while an in-memory SQL database did enable simplifying the queries, the extra coding required to get the data in and out of the database would have eaten up all the savings.

Tutorial Pages:
» Borrowing a data model can simplify development and enhance performance
» No database needed
» My kingdom for a data model
» XQuery to the rescue
» Summary
» Resources


First published by IBM developerWorks

Related Tutorials: » All about JAXP, Part 1 » Load List Values for Improved Efficiency » 2 Ways To Implement Session Tracking » A Simple Way to Read an XML File in Java » Develop Aspect-Oriented Java Applications with Eclipse and AJDT » Java Validation With Dynamic Proxies