At the department we have been analyzing some transaction data for some time. We got a new dataset with lots of transactions. Once you need interpurchase (IPT) times, posix is quite useful, as you can easily difference transactions to generate IPTs.
So the data is on a daily basis. However, some IPTs appeared to be decimal:
30.04167 days?? Well, of course R is right. So what's the reason? Marco (thx!) quickly gave me hint: daylight savings time. In my timezone there is daylight savings time. By defaukt strptime() uses the current time zone. Switching to UTC solved the issue.