Can waldo detect / highlight sort order differences?

[Moohan]

Occasionally, I'll compare two largeish tibbles where the only difference will be the sort order (although I don't know that for certain going in, hence the use of waldo). I think, related to #142, this is quite slow but also the output when it does appear is not very informative. I've just come to recognise these types of results where nothing matches as a likely out-of-order problem, and I just re-run waldo but first sort both data sets into what I assume will be a sensible order.

It would be great if waldo could somehow highlight the sort order differences: e.g. old sort order is likely x, y, z but the new sort order is likely x, z, y I guess that's probably a crazy complicated problem to solve though!

An alternative stand-in suggestion would for waldo to first check the head and tail of the dataset (first and last 20 records for instance) and if there's no or almost no agreement there, just give up and report those results, rather than trying to compute and report results for the entire data set.