Structured query language (SQL) has been called overly simplistic since its original description, but it has been enhanced and elaborated upon over the years to become a very capable data programming language. The authors propose an additional extension to SQL, to allow users to manipulate their data in a multi-dimensional fashion, equivalent to a spreadsheet.

This paper describes the proposed extension, and clarifies it with easily readable examples. It then describes some of the execution issues, and proposes algorithms to minimize the impact of those issues. Experimental data is provided to prove that the algorithms have a positive impact on the query execution time.

Business users frequently analyze their data in a multi-dimensional table, and most analytical tools are based on this concept. This extension would allow data analysts to convert a set of one-dimensional tables into one multi-dimensional table, for input to the analysis process. I see only one potential issue: execution performance. Optimization of that problem will require additional experimentation, and may require tuning, both of the creation process (the relational database engine with this extension built into it) and the receiving process (typically an online analytical processing (OLAP) tool that expects to be provided with a set of one-dimensional tables).

With commercial implementations containing millions of rows of data, converting from one-dimensional tables into a multi-dimensional, sparsely populated table is an important problem. I recommend that readers consider how this could impact their proposed solutions, or problems that might now be tractable that were not tractable with conventional approaches.