So what can we conclude? Well, with the drivers here I focused primarily on ease-of-use. There are other factors that need to be considered, as well. Do they support connection pooling, for example? Do they cache? What about pulling in large amounts of data? (Hint: Most of the better drivers for most of the popular languages support cursors, so you don’t have to pull all the data in at once.) Those are factors you’ll need to investigate as you choose a driver for the language and database you’re using. But in general, virtually all the popular languages today, including Java, PHP, Python, PERL, and even C++, have nice libraries that make database programming far easier than it used to be.
via SQL vs. NoSQL: Which Is Better?.
MongoDB does great with large complex structures that are typically read in individually, while the large relational databases do well when I’m processing huge amounts of data. And no, my clients’ data needs are nowhere near as big as Google, so we don’t encounter any performance and scalability problems.