<pushes glasses up nose>
Actually, the Correct Answer is a bloom filter.
(And, yes, we had math in the early 2000s.)
Snark aside, I'm frustrated for the author. Her completely-reasonable schema wasn't "terrible" (even in archaic MySQL)—it just needed an index.
There's always more than one way to do something. It's a folly of the less experienced to think that there's only One Correct Way, and it discourages teammates when there's a threat of labeling a solution as "terrible."
Not to say there aren't infinite terrible approaches to any give problem: but the way you guide someone to detect why a give solution may not be optimal, and how you iterate to something better, is how you grow your team.
To be fair, the “correct way” to do databases at that time was to use third normal form. Putting indices on string columns would have been considered a hack, much like using MySQL was.
3NF is largely orthogonal to indices on string columns, though. If you need to look something up by string fast, there's no avoiding an index of strings somewhere.
Plenty of databases at the time had pretty lame string indexing because it simply wasn't something people relied on if they wanted performance, and memory sayeth mysql's were both not particularly efficient and had some nasty technical limitations.
On the mysql she was using, breaking things out so it only needed to index ints was almost certainly a much better idea. On anything I'd be deploying to today, I'd start by throwing a compound index at the thing with the expectation that'd probably get me to Good Enough.
When has ever been putting indices on string columns "a hack"? If you're looking for that string, which seems to have been the case here in particular, you need that index.
(And, yes, we had math in the early 2000s.)
Snark aside, I'm frustrated for the author. Her completely-reasonable schema wasn't "terrible" (even in archaic MySQL)—it just needed an index.
There's always more than one way to do something. It's a folly of the less experienced to think that there's only One Correct Way, and it discourages teammates when there's a threat of labeling a solution as "terrible."
Not to say there aren't infinite terrible approaches to any give problem: but the way you guide someone to detect why a give solution may not be optimal, and how you iterate to something better, is how you grow your team.