The export process fetches metadata and data from a relational database and persists it as XML files.
It does the following:
Extracts database metadata information (table- and columns names, types, etc.) using Java code (JDBC) and serializes it as XML.
Uses this metadata XML as an input and creates an Ant export configuration for DBUnit. Code: XSLT.
Runs DBUnit to export the data and serializes it as XML (relational model), since - as mentioned earlier - DBUnit has Apache Ant integration. Code: XML.
DBUnit is an open source tool to export and import database data to and from XML datasets. It's database vendor agnostic, has a connector for most widely used databases. It has a streaming mode to deal with big data. In our configuration, we exported each table into a single XML file. The largest XML output file so far was 43 GB.
Example 4. XML data using the relational model - from DBUnit:
<table name="employee">
<column>employee-id</column>
<column>first-name</column>
<column>second-name</column>
<row>
<value>e1</value>
<value>Ola</value>
<value>Nordmann</value>
</row>
<row>...</row>
</table>