SIGUSR2 home apg(7) colophon feed

% A Spreadsheet is Not a Database % spreadsheets,internet-corrections, #pinned % 2013-09-24

[Thinking of Databases as an Excel File][1] is too simplistic a metaphor and misses the main point that differentiates the two.

But, Justin isn’t completely wrong. Sure, a spreadsheet can represent rows and columns of data (i.e. tabular). And, some spreadsheet packages actually go so far as to allow you to have multiple “sheets”, which, with a bit of stretching and (perhaps) misuse can even be related to another table for use in a custom input (e.g. a drop down list) or who knows what. There’s also this concept of [pivot tables][2]–the equivalent in a relational database being an aggregate query.1

Why is his metaphor too simplistic? It relates[^2] to the fact that a spreadsheet’s data is “live.” It’s a living thing, reacting to changes in input, changes in formulas, newly added formulas, etc.. The difference between a relational database and a spreadsheet is simple. The spreadsheet is a live, running program, and the tables in a relational database are nothing more than rotting bits.


  1. Hopefully you don’t notice, but it’s possible in some RDBMSes to create what are known as views (or materialized views, or both). With these, you could essentially define a query (or a procedure) that computed the same data as could be found in your spreadsheet, in a sort of “live” fashion. Updates either occur on some timer, or when you request it via a query. 

    [^2]: No pun intended

    [1]: http://justin.abrah.ms/mentoring/databases-as-an-excel-file.html

    [2]: http://en.wikipedia.org/wiki/Pivot_table

    [3]: #note-view

    [4]: #note-pun

    [5]: #return-view

    [6]: #return-pun