SQL Standard
SQL:2023 Standard
As of June 2023 the new standard for SQL had been released, by the International Organization for Standards. Although SQL language has been around since the 1970s, this is the 11th release or update to the SQL standard. The first version of the standard was back in 1986 and the base-level standard for all databases was released in 1992 and referred to as SQL-92. That’s more than 30 years ago, and some databases still don’t contain all the base features in that standard, although they aim to achieve this sometime in the future.
| Year | Name | Alias | New Features or Enhancements |
|---|---|---|---|
| 1996 | SQL-86 | SQL-87 | This is the first version of the SQL standard by ANSI. Transactions, CREATE, Read, Update and Delete |
| 1989 | SQL-89 | This version includes minor revisions that added integrity constraints. | |
| 1992 | SQL-92 | SQL2 | This version includes major revisions on the SQL Language. Considered the base version of SQL. Many database systems, including NoSQL databases, use this standard for the base specification language, with varying degrees of implementation. |
| 1999 | SQL:1999 | SQL3 | This version introduces many new features such as regex matching, triggers, object-oriented features, OLAP capabilities and User Defined Types. The BOOLEAN data type was introduced but it took some time for all the major databases to support it. |
| 2003 | SQL:2003 | This version contained minor modifications to SQL:1999. SQL:2003 introduces new features such as window functions, columns with auto-generated values, identity specification and the addition of XML. | |
| 2006 | SQL:2006 | This version defines ways of importing, storing and manipulating XML data. Use of XQuery to query data in XML format. | |
| 2008 | SQL:2008 | This version includes major revisions to the SQL Language. Considered the base version of SQL. Many database systems, including NoSQL databases, use this standard for the base specification language, with varying degrees of implementation. | |
| 2011 | SQL:2011 | This version adds enhancements for window functions and FETCH clause, and Temporal data | |
| 2016 | SQL:2016 | This version adds various functions to work with JSON data and Polymorphic Table functions. | |
| 2019 | SQL:2019 | This version specifies muti-dimensional arrays data type. | |
| 2023 | SQL:2023 | This version contains minor updates to SQL functions to bring them in-line with how databases have implemented them. New JSON updates to include a new JSON data type with simpler dot notation processing. The main new addition to the standard is Property Graph Query (PDQ) which defines ways for the SQL language to represent property graphs and to interact with them. |
The Property Graph Query (PGQ) new features have been added as a new section or part of the standard and can be found labelled as Part-16: Property Graph Queries (SQL/PGQ). You can purchase the document for €191. Or you can read and scroll through the preview here.
For the other SQL updates, these updates were to reflect how the various (mainstream) database vendors (PostgreSQL, MySQL, Oracle, SQL Server, have implemented various functions. The standard is catching up with what is happening across the industry. This can be seen in some of the earlier releases
The following blog posts give a good overview of the SQL changes in the SQL:2023 standard.
Although SQL:2023 has been released there are some discussions about the next release of the standard. Although SQL:PGQ has been introduced, it also looks like (from various reports), some parts of SQL:PGQ were dropped and not included. More work is needed on these elements and will be included in the next release.
Also for the next release, there will be more JSON functionality included, with a particular focus on JSON Schemas. Yes, you read that correctly. There is a realisation that a schema is a good thing and when JSON objects can consist of up to 80% meta-data, a schema will have significant benefits for storage, retrieval and processing.
They are also looking at how to incorporate Streaming data. We can see some examples of this kind of processing in other languages and tools (Spark SQL, etc)
The SQL Standard is still being actively developed and we should see another update in a few years time.
Ora-lytics 