Focusing on open APIs for enterprise applications

Open Web Magazine

Subscribe to Open Web Magazine: eMailAlertsEmail Alerts newslettersWeekly Newsletters
Get Open Web Magazine: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn

Open Web Authors: Jnan Dash, Jayaram Krishnaswamy, Bob Gourley, Kevin Benedict, Pat Romanski

Related Topics: XML Magazine, Java Developer Magazine, Open Web Magazine, PHP Developer's Journal

XML: Article

Open Source Database Special Feature: An Introduction to Berkeley DB XML

Basic concepts, the shell commands, and beyond

In this article I am going to introduce you to the latest version of the Berkeley DB XML, version 2.2.8. Berkeley DB XML (BDB XML) is built on top of the well-known Berkeley Database (BDB). BDB XML is an open source, native XML database. Like its ancestor, BDB, it's an embedded database. It provides APIs for the Java, C++, Perl, Python, PHP, and Tcl languages. It supports the popular XML query languages XQuery and XPath 2.0. I will show you how to use BDB XML in two ways. This month I will introduce the BDB XML shell, and next month we will explore using BDB XML with Java. BDB XML has a lot of features, and I will try to cover the most important ones.

What's an Embedded Database?
Some of you may be familiar with embedded databases. An embedded database runs within another program. It is not a stand-alone server such as Oracle, DB2, or eXist. It is the programmer's responsibility to invoke the proper API for the database. For example, although Berkeley DB XML has transaction support, in order to activate it, API calls have to be made by the programmer. Many popular relational and native XML databases run in client-server mode. Oracle, DB2, and eXist run as client-server applications, while database server runs as a stand-alone application and many clients can connect to it. Clients and server programs run in different address spaces, probably even on different machines. A connection can be established between them using the JDBC, or ODBC protocols. In an embedded database such as BDB XML, this is not the case; clients and server both run in the same address space, and they are actually the same program. The client manages the Berkeley DB XML through API calls. Next month I will give some Java examples, which will clarify what an embedded database is.

Installing BDB XML on the Windows operating system is very easy. BDB XML comes with an installer for Windows. If you are using a UNIX-like operating system you have to build BDB XML from the source code. See "Building Berkeley DB XML for UNIX/POSIX systems" for detailed build instructions. I installed BDB XML on my laptop, a 1.4GHz Pentium M processor running on Windows XP Professional with 512Mb main memory, without any problems. The installer will automatically set the necessary Path and Classpath environment variables.

Using dbxml Shell
The shell is located under the bin directory of the BDB XML installation directory. It should be in your Path, so you can run it from any location at the Windows command line.

Invoke the shell by typing "dbxml," and type "quit" to terminate the shell at any time. In order to list all available commands, type "help." To obtain detailed information about a command, type help and the command name. For example "help createContainer" will display usage information of the createContainer command.

First of all we have to create a container in which to store our data. A container in BDB XML is similar to a folder. Many XML files can reside in a container. In this article I am going to use XBench sample XML files and queries. You can download the XML sample files and the XML Schema from

Let's create a container called xbench.dbxml. You can, of course, name it anything you want.

dbxml> createContainer xbench.dbxml
Creating node storage container with nodes indexed

The user can choose between two storage types: Wholedoc or Node. The default is Node type storage. If you choose Wholedoc storage, the XML files will be stored as they are with all of the white space preserved. Node storage performs better. The BDB XML documentation suggests not using Wholedoc storage for documents bigger than 1MB.

Now we are ready to put some XML data into xbench.dbxml container. In BDB XML the opened container is the default container. We have to open xbench.dbxml before working with it.

dbxml> openContainer xbench.dbxml

Actually, dbxml shell commands on Windows are case insensitive, so it's possible to write opencontainer instead of openContainer, but I am going to follow the established naming conventions here.

dbxml> putDocument sample_10 C:\dictionary10.xml f
Document added, name = sample_10

The putDocument command takes three arguments: the first argument is the unique identifier of the document, the second argument is file name, and the third is either f or s. I chose the name sample_10 as the unique identification of this XML document. C:\dictionary10.xml is the file name, including the path. Finally, the last argument f states that it's a file. Instead of providing a filename it's possible to provide XML data itself within quotes. There are several examples in the "Introduction to Berkeley DB XML" document, which uses an XML string instead of a file name.

As mentioned earlier, BDB XML supports XQuery and XPath 2.0 languages. There will be several examples for each language (see Listings 1-4). BDB XML found 28 results for this query. Note that in order to retrieve the results we have to use "print" command.

BDB XML will produce more explanation about what it does when the setVerbose feature is turned on. This command takes two numeric arguments: level and category. Using this command, let's get more information regarding the XQuery Example2 (see Listing 5).

No indices are used in this query. The execution time is around 48.5 seconds. Execution time is very high for this query for two reasons: this is not a trivial query, and there are no indices created yet. I will conceptually explain the possible execution plan for this query:

For each $entry as /dictionary/e
  For each descendant $desc of $entry
   For each value $val of $desc
   Check if 'hockey' is a substring
    of $val
   return $entry/hwg hw

There are three for loops in this algorithm. At the heart of the three loops there is a very costly substring (contains) operation. Considering that no indices have been created, execution time is pretty good. Creating indices improves the query performance dramatically (see Listings 6-7). (You can find more information about indices in the Indexing section of this article.)

More Stories By Selim Mimaroglu

Selim Mimaroglu is a PhD candidate in computer science at the University of Massachusetts in Boston. He holds an MS in computer science from that school and has a BS in electrical engineering.

Comments (5)

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.