Correct option is
As a non-relational database, there are some things that Hadoop cannot do. Hadoop cannot, for instance, be able to validate dates, account balances, and other input records, the way that a MySQL or other relational database can. This is a major reason as to why Hadoop is not considered to be a replacement for a traditional relational database; though it can store large volumes of data, it requires another layer to actually interpret and verify this data. Hadoop is able to comb through data quickly, but it does so at the expense of the relationships between this data.
Hadoop also requires fairly extensive resources for some operations, such as joins. But the interesting thing about Hadoop is that many of its weaknesses are lost when it is combined with relational technology such as SQL.
By layering Hadoop onto a relational database structure, the weaknesses of both systems are resolved; the system can crunch large amounts of data quickly, but can also relate the data and verify it as needed.
Not only is Hadoop not sufficient for replacing RDBMS, but it’s not what it truly is meant to do. Hadoop is designed to make it easier to use a traditional, relational database, by speeding up operations that directly relate to large data sets. Though it may have many benefits in raw data fields, Hadoop cannot (and usually has not) replace a data warehouse. When mixed with relational databases. however, it creates a powerful and versatile solution.