One way to modify a type tree to increase performance is to:

Mapping unused elements to =NONE still requires the parser to individually locate and evaluate each element before discarding it, providing no parsing skip benefit.

When contiguous elements are not mapped, merging them into a single blob lets the parser treat the entire group as one opaque unit to skip rather than individually parsing and evaluating each unused element. This reduces the total number of parsing operations the engine must perform, directly lowering CPU time and improving map execution speed. The technique is a recognized type tree optimization that avoids unnecessary traversal of unused fields.

Moving unused elements to a nonUsed category is not a recognized optimization technique and does not reduce parser work because the engine must still traverse those elements.

Changing numeric elements to alphanumeric does not reduce parse overhead and can increase processing time by requiring additional type conversion logic.