SQL vs NoSQL: which technology will dominate in the Big Data era?

In previous articles I summarised the features of SQL and NoSQL as technologies that cover the need to structure and manage large volumes of data.

To summarize what I previously wrote, the great difference between both technologies is that SQL offers a high degree of structure by offering relational database management systems. On the other hand NoSQL do not use tables and their main language is not SQL.

An important question remains to be answered: which one do you think will chosen by companies in the near future? Let’s go over the differences between both types of technology in two regards that experts consider important: Flexibility and Scalability.

Comparing SQL and NoSQL: flexibility and scalability. 

Flexibility: SQL uses relational databases so each entry in the database, that is, each row, complies with a fixed scheme. This means that the number of columns and column names must be defined before creating the tables and entering data into the database. Each row should contain data for each column unless columns are configured specifically to allow null values, this is, absence of data.

This does not happen with the technologies encompassed by NoSQL. The schemas are not fixed but dynamic. Data can be added at any time and without a predefined way in which you will work. This is a very important functionality today as it allows rapid changes in the databases.

Scalability : Traditional SQL systems are vertically scalable. This means that databases are usually hosted in a single server to ensure a good performance in operations involving multiple tables. Therefore, as the volume of data increases, so does the needed server storage capacity. This ends up being quite expensive if over time we incorporate more entries to our databases. In NoSQL, scalability is horizontal. Therefore, instead of adding more storage capacity on a single server, more servers are added.

Although this could also be done with SQL, it is not optimal because it requires high-cost tasks and time to make all servers work as one single database.

NoSQL databases can expand their data between multiple servers in an automatic way. These servers can be quite cheap compared to a high capacity server.

So is NoSQL going to replace SQL?

This really is a difficult question to answer considering the current state of the tech industry. We live in an age in which the volume of data is excessive and in continuous change. This is a self-evident challenge for relational database management systems because their highly fixed structure poses problems  to manage all that information.

However, companies have invested lots of time and resources in building their databases and those who manage them have spent lots of time mastering SQL. 

Therefore, in my opinion, the replacement of SQL by NoSQL is not a change that will happen overnight. RDMS licenses purchases have decreased significantly in recent years and NoSQL technologies have become more popular. However, the use of open source SQL solutions has experimented a great growth over the recent years, especially MySQL. Here organizational culture also matters; companies have worked with a relational databases for decades and now that they have access to free stable technology it makes sense to stick to your current working paradigm.

Another reason why NoSQL is still far from replacing SQL is that, according to Data Science experts, NoSQL analytical tools are neither accessible nor easy to use and they still have a long way to mature.

For these reasons it is difficult to affirm that we are approaching a reality without SQL. Rather, it is more likely that both technologies coexist and are used complementarily depending on the analysis strategy used.


Despite NoSQL technologies are reaching maturity, relational databases may still be the best solution to certain problems. In many occasions we will be able to to choose according to our needs. If the data you need to analyze is bulky and changes fast, NoSQL is the best solution. But if the data you handle is stable by nature and can be well structured in tables, SQL technologies will allow you to make fast powerful queries. In short, I think there is no one-size-fits-all solution. What is clear is that both SQL and NoSQL technologies promise to be great analysis  tools in the era of Big Data

More info in:

Data science skills: Is NoSQL better than SQL?

NoSQL keeps rising, but relational databases still dominate big data