In OD data it is common to have many 'oneway' flows from "A to B" and "B to A".
It can be useful to group these an have a single ID that represents pairs of IDs
with or without directionality, so they contain 'twoway' or bi-directional values.
od_id*
functions take two vectors of equal length and return a vector of IDs,
which are unique for each combination but the same for twoway flows.
This function superseeds od_id_order as it is faster on large datasets