Solutions

MarkLogic Data Hub Service

Fast data integration + improved data governance and security, with no infrastructure to buy or manage.

Learn More

Learn

Stay On Top Of Everything MarkLogic

Be the first to know! News, product information, and events delivered straight to your inbox.

Sign Me Up

Community

MarkLogic World 2019

Learn how to simplify data integration & build innovative applications. Join us in Washington D.C. May 14-15!

Find Out More

Company

Stay On Top Of Everything MarkLogic

Be the first to know! News, product information, and events delivered straight to your inbox.

Sign Me Up

Looking at a Mail Message

Take a minute to look at the XML format returned by the above query. It's the genuine internal document format used by MarkMail to represent emails. You may be thinking, hey, emails aren't usually written in the XML format. That's right, they're usually sent over the wire in a format known as RFC 822, and persisted on disk in a variety of formats like mbox, maildir, or PST. MarkMail reads email in these various formats and converts them into this singular XML model.

The <message> root element has several attributes that provide metadata about the email: its list, date, MarkMail id, mail message id, etc. Below that are all the various mail headers are gathered under a <headers> element. A handful of important headers (from, to, cc, subject) have attributes on them which hold strings extracted from the raw header value.

The <body> element holds the actual text. Each paragraph gets its own <para> element, including a quote depth. The initial greeting and footer (if any) are detected during the load and identified with corresponding elements, which makes it easy to exclude them from searches. Any email addresses and URLs in the body are tagged. That way, during display, the emails can be obfuscated and URLs rendered as hyperlinks. (You can see we obfuscated the email addresses in the source here so spammers can't use this site to gather addresses. In the real MarkMail database the emails are unobfuscated in the source but obfuscated on display and revealed only after the user solves a CAPTCHA. That's the only change between this data and the live site.)

What's most interesting about this data model is how simple it is. It enables a robust, feature-rich site but the model fits on the back of a napkin. An email is one logical item, represented by one physical item in the database. Try to imagine how you'd model, retrieve, and query this kind of textual, hierarchical, irregular data using the relational model.

In the above query we gave you a document identifier to use. If we weren't around giving you ids, what would you do? Well, you might just ask for a message at random. Here's how you do that:

This query returns the first document in the database (order is essentially random, though it is stable and doesn't change between executions). The first message happens to be written in Russian. That's fine; MarkLogic supports storage and searching of all languages whose text can be represented in XML, and provides advanced support (tokenization, stemming, and collation) for more than a dozen languages, including Russian.

If you look at the message's id attribute and do some copy/paste work you can derive a URL to see this mail on MarkMail. In fact, here's a query to generate that link for you:

Pro tip: You can drop the leading doc() and you get the same result. Any rooted XPath is assumed to query across all documents. Let's use that shortcut and ask for ten random emails to explore:

Lay of the Land

Drilling in with XPath

Stack Overflow iconStack Overflow: Get the most useful answers to questions from the MarkLogic community, or ask your own question.

Comments

The commenting feature on this page is enabled by a third party. Comments posted to this page are publicly visible.