Syncing up iterators with gaps

[Chris Angelico]

On Thu, Sep 29, 2016 at 5:10 AM, Tim Chase
<python.list at tim.thechases.com> wrote:
> And I'd like to do something like
>
>   for common_key, d1, d2, d3 in magic_happens_here(data1, data2, data3):
>     for row in d1:
>       process_a(common_key, row)
>     for thing in d2:
>       process_b(common_key, row)
>     for thing in d3:
>       process_c(common_key, row)

Assuming that the keys are totally ordered and the data sets are
sorted, this should work:

def magic_happens_here(*iters):
    iters = [iter(it) for it in iters]
    nexts = [next(it, (None,)) for it in iters]
    while "moar stuff":
        try: common_key = min(row[0] for row in nexts if row[0])
        except ValueError: break # No moar stuff
        outputs = [common_key]
        for i in range(len(nexts)): # code smell, sorry
            output = []
            while nexts[i][0] == common_key:
                output.append(nexts[i])
                nexts[i] = next(iters[i], (None,))
            outputs.append(output)
        yield outputs

Basically, it takes the lowest available key, then takes everything of
that key and yields it as a unit.

Code not tested. Use at your own risk.

ChrisA