You can still think of data as 'rows' with MongoDB - objects belong to a collection, which is analogous to a table - and you can query and select only specific fields.
The advantage is that you aren't nearly as restricted in the kinds of data you can build. SQL often forces you to split logically related data across several tables and wrestle with complicated joins, simply because a row can't contain sub-arrays or hashes.
Of course, there's no silver bullet - some times you DO want to group together logically different data in a same query. But I've found that in general, I'm not fighting the DB as much when building stuff using MongoDB.
Of course, there's no silver bullet - some times you DO want to group together logically different data in a same query. But I've found that in general, I'm not fighting the DB as much when building stuff using MongoDB.