I'd argue the example given in the article is bad not because it doesn't repeat itself, but because it only concats strings. Using more advanced language features, or perhaps a ORM, I think you could easily come up with a solution that doesn't involve repeating parts of SQL queries but is still readable.
The anti-pattern here is applying DRY to text rather than intents. The identical text may appear in a different source file with a different intent, say a tax calculation for a region. It may very well be that two regions have the same formula, but it's obvious that you shouldn't DRY it.
The example used in the post had no real-world intent. The intent was to blindly apply DRY which doesn't count as one.