SQL has a couple legitimate limitations as a query language, largely because we expect databases to do so much more today than when it was designed. I've been both a heavy user of SQL and implementor of SQL databases focused on data models like graph, spatial, et al, so I feel the pain from both directions.
Some useful things are difficult to effectively express in SQL because it was not designed to make them expressible. Adding support after the fact introduced a lot of compromises and complexity; you can tell which parts of SQL are second-class citizens. SQL would have been designed differently if its original scope had been broader and it is an inelegant mess for some types of data models. You can make it work but it isn't pleasant. Attempts to address these gaps explain much of why SQL databases all have their own non-standard SQL dialect, and it has made SQL effectively non-portable across databases.
In principle, a more modern query language could elegantly address the much broader scope of how we use databases today in a standardized way, instead of the zoo of non-standard extensions and mismatched parts that are grafted onto specific SQL implementations.
Some useful things are difficult to effectively express in SQL because it was not designed to make them expressible. Adding support after the fact introduced a lot of compromises and complexity; you can tell which parts of SQL are second-class citizens. SQL would have been designed differently if its original scope had been broader and it is an inelegant mess for some types of data models. You can make it work but it isn't pleasant. Attempts to address these gaps explain much of why SQL databases all have their own non-standard SQL dialect, and it has made SQL effectively non-portable across databases.
In principle, a more modern query language could elegantly address the much broader scope of how we use databases today in a standardized way, instead of the zoo of non-standard extensions and mismatched parts that are grafted onto specific SQL implementations.