According to Thomas Zahn, Co-Founder and CEO of 3T Software Labs, coming to MongoDB from a relational database background is a relief. In this recent article from the 3T Software Labs Blog, he points to the flexibility and dynamic nature of MongoDB as big selling points, as well as the JSON-based query language. However, that's not to say that MongoDB is completely free of cryptic design choices that may leave developers puzzled.
In particular, Zahn's article focuses on MongoDB's $type query operator, which allows queries to filter by a certain types (only return Strings, for example). However, when it comes to Arrays, $type is a little strange: rather than returning fields that contain Arrays, it searches the individual elements of a field's Array for Arrays. In other words, an $type query for Arrays only finds two-dimensional Arrays - not exactly intuitive.
It's not that Zahn is out to smear the name of MongoDB - as is made clear at the start of his article, he's pretty fond of it - but some problems just seem so simple to fix:
Overall, I think it would be really great to have a dedicated $isArray query operator. The peculiar (and at times inconsistent) way arrays are handled by the search engine currently make it unnecessarily complex to guard one’s update queries against the tricky intricacies of arrays.
Check out Zahn's full article for details on how to avoid Array mishaps with $type in MongoDB.