Query By Excel

Spreadsheets, and MS Excel in particular, are established analysis tools. They offer an attractive user interface, provide an easy to use computational model, and offer substantial interactivity for what-if analysis. However, as opposed to RDBMS, spreadsheets do not provide a central repository hence they do not provide shareability of models built in Excel and lead to proliferation of multiple copies of the same spreadsheet. Furthermore, spreadsheets do not offer scalable computation, for example, they lack parallelization. To address the shareability, and scalability problems, we propose to automatically translate Excel computation into SQL. An analyst can import the data from a relational system, define computation over it using familiar Excel formulas and then translate and store it as a relational SQL view over the imported data. The Excel computation is then performed by the relational system. To edit the model, the analyst can bring the model back to Excel, modify it in Excel and store it back as an SQL View. We refer to this system as Query by Excel, QBX in short.