XML Documents and When to Use XML vs. a Relational Database

An XML Document With Two People

<?xml version="1.0" encoding="UTF-8"?>
<people>
    <person id="1">
        <firstName>John</firstName>
        <lastName>Doe</lastName>
        <email>johndoe@example.com</email>
        <phone>585-555-0142</phone>
    </person>
    <person id="2">
        <firstName>Jane</firstName>
        <lastName>Doe</lastName>
        <email>janedoe@example.com</email>
        <phone>585-555-0198</phone>
    </person>
</people>

An XML Document With One Person and Multiple Addresses

<?xml version="1.0" encoding="UTF-8"?>
<person id="3">
    <firstName>John</firstName>
    <lastName>Doe</lastName>
    <email>johndoe@example.com</email>
    <addresses>
        <address type="home">
            <street>412 Monroe Avenue</street>
            <city>Rochester</city>
            <state>NY</state>
            <postalCode>14607</postalCode>
            <country>USA</country>
        </address>
        <address type="work">
            <street>1800 Main Street</street>
            <suite>Suite 300</suite>
            <city>Buffalo</city>
            <state>NY</state>
            <postalCode>14202</postalCode>
            <country>USA</country>
        </address>
        <address type="billing">
            <street>77 Lakeshore Drive</street>
            <city>Rochester</city>
            <state>NY</state>
            <postalCode>14468</postalCode>
            <country>USA</country>
        </address>
    </addresses>
</person>

Notice that the second document nests the addresses inside the person. XML handles this naturally because it is a hierarchical (tree-shaped) format, and a person can have one address or ten without changing the structure.

When You Would Use XML

1. Exchanging data between two different systems. XML is plain text and self-describing, so a Java application can send an XML file to a Python application and both sides can read it. This is why formats like SOAP web services, RSS feeds, and health-care’s HL7 standard use XML. The tags travel with the data, so the receiving system does not need a copy of your database schema to understand what it received.

2. Documents with mixed content and irregular structure. Think of a book, a legal contract, or a scientific paper. Some chapters have footnotes, some have images, some have neither. XML lets each element carry whatever child elements it needs. Trying to force that into fixed table columns would leave you with dozens of mostly-empty fields.

3. Configuration files. Files like Maven’s pom.xml, Android’s AndroidManifest.xml, or a Tomcat web.xml are read once when a program starts. They need to be human-readable and editable in any text editor, and they are naturally nested (a project contains dependencies, a dependency has a version). A database would be overkill for something a program reads a single time at startup.

When You Would Use a Relational Database

1. An application with many users reading and writing at the same time. An online store’s inventory, a bank’s account balances, or a university’s registration system all need transactions. A relational database guarantees ACID properties, so two people cannot buy the last item at the same moment or withdraw the same $100 twice. An XML file has no locking, no transactions, and no way to safely handle concurrent writes.

2. Querying and reporting across large amounts of data. If you want “all customers in New York who ordered more than $500 last quarter, sorted by total spend,” SQL does that in one indexed query across millions of rows. Searching the equivalent XML would mean parsing the whole file into memory and walking the tree every single time.

3. Data with lots of relationships that must stay consistent. Students enroll in courses, courses have instructors, instructors belong to departments. A relational database stores each fact once and links the tables with foreign keys, so updating a department name updates it everywhere. Foreign key constraints also block bad data, such as an enrollment pointing to a course that does not exist. In XML you would repeat the same information in many places, and nothing stops those copies from drifting apart.

Summary

Consideration	XML	Relational Database
Shape of data	Hierarchical / nested	Tabular rows and columns
Best at	Moving and storing documents	Storing and querying live data
Structure	Flexible, varies per record	Fixed schema, enforced
Concurrent users	Not supported	Fully supported with transactions
Searching	Parse the whole file	Indexed SQL queries

The short version: use XML when data needs to travel between systems or be stored as a self-contained document, and use a relational database when data needs to live somewhere, be updated by many people, and be queried quickly.