BytesIn bug fixed 
Changed the reporting script so that it prints out each IP's bytesIn and bytesOut as well as the TotalBytes. New, sexy invocation:
$ ./totals -a '2009-04-01 00:00:00' -b '2009-05-01 00:00:00' | nl | less
This prints "ranking" numbers for each line so I can see, for example, my system was #15 for total bytes
Noticed that for some systems, the reported bytes in was 0, when this clearly wasn't the case. After much fumbling, found a pretty stupid error on my part in the import script. I'm surprised that the bytesIn values were ever anything BUT zero… which of course means that all the TotalBytes values were wrong.
Fixing this means that I have to re-run the import. This might take a day or so, but it is running now.