|
|
|
|
|
by contravariant
2076 days ago
|
|
You're arguing about capability, whereas the post is arguing about the grammar itself. What you're describing is how to build a separate domain specific language that compiles to SQL (which there are quite a few of; e.g. C#'s Linq). Quite a few query languages are equally capable (including, surprisingly, quite a few so-called graph languages, provided the SQL dialect provides a transitive closure), but SQL as a language has some undesirable properties (most of which are trade-offs for the fact that basic SQL is very easy to parse). |
|