What is the difference between SQL and NoSQL databases?

SQL databases are structured and use a predefined schema, while NoSQL databases are unstructured, scalable and have a dynamic schema for unstructured data.

SQL (Structured Query Language) databases, also known as relational databases, have been a predominant choice for database management systems since the 1970s. They are based on a relational model, where data is stored in tables and the relationship between these data is also stored in tables. SQL databases use a predefined schema, which means you have to know in advance the structure of the data you are working with. They are best suited for complex queries as they allow joining tables and performing operations on them. Examples of SQL databases include MySQL, Oracle, and PostgreSQL.

On the other hand, NoSQL (Not Only SQL) databases emerged to handle the shortcomings of SQL databases, particularly concerning large data sets and the need for agility, efficiency, and scalability. NoSQL databases are unstructured, meaning they do not require a fixed schema and the data can be stored in many ways including key-value pairs, wide-column stores, graph databases, or document databases. This flexibility makes NoSQL databases ideal for semi-structured or unstructured data. They are highly scalable and allow for easy updates and changes to the database. Examples of NoSQL databases include MongoDB, Cassandra, and Couchbase.

In terms of ACID (Atomicity, Consistency, Isolation, Durability) properties, which are a set of properties that guarantee reliable processing of database transactions, SQL databases are strong in all four. NoSQL databases, however, may not provide full ACID properties but instead offer "eventual consistency" in which database changes are propagated to all nodes "eventually" (i.e., in a certain period of time).

In summary, the choice between SQL and NoSQL databases depends largely on the specific requirements of a project. SQL databases are typically used when dealing with structured data and complex queries, while NoSQL databases are preferred for large data sets and when data structure is not clearly defined or subject to change.