As a rookie to databases I have done some researching into SQL and NoSQL. There have been numerous articles and opinions to sift through and learn from. Reflecting on a few articles and opinions, the only definitive conclusion is that both SQL and NoSQL are beneficial databases.
Not that SQL and NoSQL are polar opposites, but they play the database game very differently.
SQL databases are relational databases, meaning that they store information about both the data and how it is related. These databases are table based, like a spreadsheet. NoSQL databases are codument based, and unlike SQL, are horizontally scalable.
Schema is the predefined acceptance and linking of data structures, allowing the access or joining of data amongst multiple tables. This is only used in SQL.
As stated in the article from news.dice by Jeff Cogswell, some argue that the joining of multiple relational databases can slow the system down when exponential data and users. He exemplifies Amazon and Google, which compile massive amounts of data and have multitude of users and now use non-relational systems to suit their needs. Additionally he offers the argument that relational databases does not map well to typical structures of complex datatypes or hierarchal data which can be problematic or difficult to manage.
Although there is a growing use of NoSQL, there is predominately more resources and support for all SQL databases and their vendors.
In reading an article and overviewing the properties of both databases, it appears that SQL databases are more stable and better suited for 'heavy duty transactional type applications.' As stated by Luke P Issac, "While you can use NoSQL for transactions purpose, it is still not comparable and stable enough in high load and for complex transactional applications." NoSQL is ideal for large amounts of data, while it appears that for the relations model for databases like SQL are the popular choice for data reliability, security, and, for the most part, efficiency.