You may have heard another pandas user mention using eval and query to speed up evaluation of expressions in pandas. While use of these functions can speed up evaluation of expressions, it cannot do it without the help of a very important library: NumExpr. Use of these functions without installing NumExpr can actually cause a performance hit. In order to understand how NumExpr is able to speed up calculations however, we need to take a deep dive into the architecture of a computer.
Computer architecture
CPUs are broken up into multiple cores where each core has a dedicated cache. ...
