I’m not sure about any performance gains or working with large datasets, but the ancient Metakit[1] was just a really pleasant relational algebra library ( ≠ SQ

I’m not sure about any performance gains or working with large datasets, but the ancient Metakit[1] was just a really pleasant relational algebra library ( ≠ SQL data model library, it could do e.g. relations as values which are difficult for row-oriented databases). I’d say that Metakit & OOMK in Tcl is strictly better than the relational part of Pandas in Python, except the documentation is somewhere between bad and nonexistent. [1]: https://git.jeelabs.org/jcw/metakit https://git.jeelabs.org/jcw/metakit