Marklogic vs HBase
HBase follows in the likes of BigTables by supporting millions of columns in a distributed environment. This works well in BigData for scaled analytics discovery. Unfortunately HBase will not successfully facilitate highly complex queries or AI/ML its hard to support these queries regardless of however many nodes available. Its also highly inefficient way of bruteforce querying.
Marklogic supports a different paradigm with a Triple store. This is lots of rows in a simplified highly normalised structure. Complex queries are broken down into lots of small queries returning results with a performance factor of >100x over bigdata.