@jamesevans
Not that it matters much but the same thing is done for the weekly bus WTT upload; every live file is recreated! In a typical week, about half of the recreated files have a slightly different file size. Not that I can talk; I do the same for the route-level landing pages on my bus WTT website!
@jamesevans , I take your point about total dataset reload but, nevertheless, the data does change day on day. I take the output from /Line/Mode/{modes}/Route each day and check for change of created/modified, for expiry (i.e. no routeSection where validFrom>ValidTo includes current timestamp), and for change of data (i.e mismatch between previous and current routeSections). This throws up a number of routes on most days, which shows me that the data is being selectively amended , albeit without any update to dateModified.
Incidentally, I think your use of validTo needs tweaking. I notice, particularly on a Friday, a number of routes with a validTo timestamp of, say 20/10/2023 00:00:00, where the actual changes don’t appear until the following day. I think your timestamps should be using 23:59:59 for the time component of validTo.
The 675 appears to have been start-dated incorrectly in the source system for schedules. This should appear on Monday automatically as there is a schedule loaded that starts on this day.
The 63 and 638 have been fixed and will appear when we release next week’s build.
The 698 is back now. With school buses (i.e. no weekend service), old schedules end on the Friday and new schedules start on the Monday. Often, this will lead to an “interregnum” between the 2.