Listen up, SQL and relational database people: The knowledge revolution has reached the SQL world and it will change it forever.
You may have heard about knowledge graphs.
Maybe you know that Google coined the term to describe the technology behind some of its platforms and features such as the information boxes displayed at the right of search results.
You may have read that companies such as Amazon, Facebook, Microsoft, JPMorgan and Bank of America have made large investments to develop their own proprietary knowledge graphs, to make “strategic use of data and extend business boundaries”.
How is this relevant to the “SQL World”?
What does this have to do with you, the SQL/relational database professional? Why should you care that there is a new world of databases that is alien to most relational database experts and users?
For the near future maybe you shouldn’t care. After all, about 80% of the database infrastructure in the world is relational, so for you SQL is a sure bet.
But wait, here’s big data, ever growing big data. Big data is complex, it has many varieties and it is difficult to manage so organizations actually use only a fraction of the data they store. Throughout enterprises there are hundreds of thousands of tables needing to be connected to each other in order to enable data scientists and analysts to derive meaningful insights using ever more complex analytics. There are machines that need to understand each other by agreeing on common meanings of data. There’s also artificial intelligence that wants to make use of big data to develop deep learning algorithms, train ML-models, etc.
All of this requires an ability to represent the connected world by describing relationships and their contextual information.
About 20 years ago the academy tried relational databases and SQL to deliver an efficient solution to these requirements. But relational databases are all about tables, columns and rows. It is difficult to represent and derive knowledge from tables, columns and rows. You need many Union and Join statements so SQL queries that represent connected data reach even hundreds of lines of code. You can’t visualize and explore tables and columns as connected data either. In addition, in SQL there’s not such a thing as an abstract concept to represent a common meaning of columns in separate tables or databases.
A gap is born
So, unable to find a suitable, efficient answer in relational databases and SQL, academia developed the Semantic Web principles, adopted as standards by the W3C to represent knowledge. They incorporated graph theory to represent connections and added semantics: ontologies of abstract concepts that describe the data in hierarchies and relationships incorporating contextual information. They developed OWL to model the ontologies, RDF as the data format and SPARQL to query them both. But OWL is complicated and SPARQL is not widely known. RDF stores require different backends and new skills must be learned to implement them. So, after 20 years from its inception, this initiative to represent knowledge, which is largely disconnected from the relational world, is very, very far from becoming ubiquitous, especially in light of the wide and still growing popularity of relational databases and SQL.
In parallel to the Semantic Web, less complicated “Labeled-Property” Graph databases (such as Neo4j and Tiger-Graph), each with its own query language, were developed by a number of companies to represent relationships without really getting into the more sophisticated semantics needed for knowledge representation. This limits their ability to share common meaning of concepts that allow machines to understand each other and limits the range of use-cases to which they can be applied (because “a little semantics goes a long way”..).
And now the semantic SQL revolution has arrived
The old idea of using SQL to enable relational databases to represent knowledge and allow users to benefit from the promise of knowledge graphs has been realized with the SQL Knowledge Graph platform that implements new ideas and architectures that were lacking 20 years ago.
The SQL Knowledge Graph platform bridges the gap between the “old” relational world and the “new” connected world of knowledge representation. For the first time, organizations can enable, visualize and query their relational databases as knowledge graphs. For the first time, any SQL-fluent professional can easily model and create knowledge graphs and query ontologies of concepts that represent the real world, on top of their existing relational databases and without moving data – in standard SQL. For the first time users can create simple SQL queries to explore and visualize relational databases as a graph of connected data. For the first time, any SQL-fluent database can be empowered with reasoning capabilities so artificial intelligence can conveniently make use of big data.
Now, any company, not just large organizations with research labs, can implement in just a matter of weeks their own knowledge graphs with the same kind of features as those developed at great cost by the technology giants. Any company can also leverage its growing investment in BI tools to query the knowledge graph using JDBC or ODBC connectors, without retraining analysts and data scientists.
Ah, let’s not forget to mention: using the SQL Knowledge Graph, SQL query length is reduced by up to 90% and there’s no need for Union and Join statements anymore.
That is what semantic SQL is all about.
Come and learn how the SQL Knowledge Graph can empower your organization to benefit from the knowledge revolution.
 Alan Morrison, PWC – “Data-centric design and the knowledge graph” – presentation 2019
 Prof. James A. Hendler – Semantic Web pioneer