If pandas gives you a full relational API with arbitrary data then in isolation it doesn’t make much of a difference. SQL is more portable so to speak, but a library like that may introduce less friction. Pragmatism is advised here.
The big idea here is to use relational logic programming to express data transformation outside of storage access. The paper „Out of the Tar Pit“ proposed this as a way to reduce accidental complexity.
The big idea here is to use relational logic programming to express data transformation outside of storage access. The paper „Out of the Tar Pit“ proposed this as a way to reduce accidental complexity.