In #rascalmpl there is no "object-relational mapper" that bridges a "impedance mismatch" because relations are a built-in general programming construct. To query is to program and to program is to query. Query syntax is comprehensions plus relational calculus.
This integration of database/graph features into PL in Rascal is very satisfying because it avoids all the accidental complexity of bridging between two worlds which are otherwise arbitrarily different.
There is only one kind of string, one kind of integer, etc, which is accidentally not the case when you are bridging a database system to a programming language. The core design principle of Rascal is to avoid all such accidental integration issues for the programmers, so
they might focus on their real challenges of software analysis and transformation applications.