I recently came across 6 Optimization Tips for Ruby MRI. One of the tips provided is to use string interpolation instead of concatenation. I set out to determine whether or not we really get any performance boost out of this, and, if so, how much.

My test code makes use of better-benchmark, my Ruby library for statistically correct benchmarking.

Here is a table summarizing the detailed results. By “append” I mean usage of the << operator; by “concatenate” I mean usage of the + operator.

The first thing to note is that the apparent performance improvement of 3.4% in using interpolation instead of appending is not statistically significant. That is, the data do not provide evidence supporting the claim that interpolation is better (or worse) than appending. However, in fairness, doing two appends makes the difference significant.

We also see that concatenation is slower than appending, as is often cited. But, again, we cannot conclude this is the case when comparing only a single concatenate versus a single append. The difference is only pronounced when doing multiple concatenations or appends in a single Ruby line.

Overall, there appears to be good reason to use interpolation by default, and append or concatenate only under special circumstances. In the end, though, you should be benchmarking larger sections of your code, such as measuring the time taken to render a web page. Doing even a hundred individual concatenations takes just a tiny fraction of a second, and a performance improvement of a few microseconds isn’t going to mean much when you have a large bottleneck elsewhere, such as in your database queries.

Related posts:

1. Concatenating row fields in PostgreSQL
2. Faster applications by using SQL over ORM
3. Ruby's grenade operator
4. Ruby: "string not matched"
5. Substrings in Ruby