Last time, I presented a couple screenshots from a recent user experience I had on the web. (If you missed it, go back and read it real quick, or the rest of this post won’t make much sense).

Thanks to everyone who responded. It was fun to see the variety of approaches that people took to solving the problem of cleaning up the “miles driven” field automatically.

For what it’s worth, the approach I would have taken would have been to use the before_validation helper in the model, as a couple of people suggested (hence the words “model validation” in the original article title). There were a few suggestions that seemed to imply that the cleanup would happen in the controller, but I don’t think that’s the right place. By making the model responsible for its data, this logic becomes more reusable and it becomes easier to attach error messages as needed.

For the curious, to just remove commas and ignore any digits after a decimal point, this will work:

miles.delete(',').to_i

but this doesn’t remove any other non-digit characters, so a value like “A123.4” will become zero. This is one of those rare moments where C# is a bit better, as Int32.Parse would come in real handy here.

But rather than try to come up with a more complicated regular expression to get rid of letters and other invalid characters, I would simply use another model validation helper to enforce a greater-than-zero value:

validates_numericality_of :miles, :greater_than => 0, :message => "must be higher than zero"

I guess my point is, I think sometimes we try to do too much. Rails is a pretty big framework and often has something built-in to make your job easier than you think.