As with many No-SQL databases, MarkLogic stores content as XML or JSON documents, although text and binary content may also be stored. While MarkLogic provides an ODBC driver and interface, using this approach requires some pre-configuration of the database and compromises your ability to extract the hierarchal structure of the stored documents.
QlikView Expressor can extract data from MarkLogic using the vendor supplied REST API. Consequently, it is possible to retrieve information without losing the structural information in the store content.
The attached PDF file describes how this approach works. Also attached is an exported Expressor Library that contains a datascript module that you can use in replicating this work.