I currently have a pipeline that originally first called filter, then partition, but I've had to combine the two closures due to a borrow-checker issue. It would be nice if there was an easier way in such cases to avoid the overhead of either building a HashMap (that in this case only has 2 keys) in the flat_map().into_group_map() case, or tupling each item with a bool in the flat_map().partition() case. Why not have a method that would flatten an array of iterators from either an array of sub-iterators per input item, or a categorizing function that returns usize (or Option<usize> to simultaneously filter), or even from an Into<usize> element 0 of a 2-tuple but then having only element 1 as the output item?
I currently have a pipeline that originally first called filter, then partition, but I've had to combine the two closures due to a borrow-checker issue. It would be nice if there was an easier way in such cases to avoid the overhead of either building a HashMap (that in this case only has 2 keys) in the flat_map().into_group_map() case, or tupling each item with a bool in the flat_map().partition() case. Why not have a method that would flatten an array of iterators from either an array of sub-iterators per input item, or a categorizing function that returns usize (or
Option<usize>to simultaneously filter), or even from anInto<usize>element 0 of a 2-tuple but then having only element 1 as the output item?