3 ms·
I am not really sure we need a 'better SQL' - SQL has been doing what SQL does in pretty much the same way for 50+/- years - if you work with relational data, l
by ejb999 5y ago
I am not really sure we need a 'better SQL' - SQL has been doing what SQL does in pretty much the same way for 50+/- years - if you work with relational data, learn SQL.
Very few technologies have had the staying power that SQL still holds.
- wruza 5y agoIs that a good argument? C also does what it does for 50+- years. If you work with code, learn C. SQL is a syntax for queries, and it is just bad at some sides. Global-only scoping, issues with common expressions, excessive roundtrips, inability to fetch hierarchical data effectively. Even syntactically it is an IDETIFICATION DIVISION-style spells, whose structure is hard to feel.
- ejb999 5y agoThis tool is just another way to write those very same SQL queries - I have yet to see any tool that purports to write the SQL for you, output SQL that is better than someone that actually understands relational data and sql can do - it is usually not even close. This isn't a new language, its a layer on top of a language that already works, and works well.
- cloverich 5y agoSQL is my least disliked technology of all the ones I use in anger. And reading one solid SQL book has served me well for 10+ years now. The only ask I would want is better errors. But even without it, I rarely spend much time debugging SQL or trying to figure out what Clever Programmer intended their code to do.
- jandrewrogers 5y agoI think the SQL DML is far from optimal but it mostly works, even though it is quite difficult to use for some cases. It will be with us for a long time. SQL DDL, on the other hand, is broken as a standard because it assumes implementation details of the database engine that were widely true when it SQL was designed but are nonsensical or increasingly divergent in modern database engine architectures e.g. the many tacit assumptions about index properties.