Only the JSON example discusses performance differences. The rest of the article is feature comparison. Of course my query will be faster on a partial index that satisfies my domain need.
It’s a strange JSON example, too – only 206 rows of 14MB apiece.
They don’t indicate whether the postgres table is using json or jsonb (the latter allows for extremely flexible indexes on the entire doc field, among other things).
Overall, confused by the aim of the article.
Yes, it’s a very thin piece, when it comes to performance differences discussion. I was expecting more data and measurements, but the article ended up being just a comparison of features.