Hello,
I am having a problem with coding a Google Transit GTFS feed and need
some help please: I've been finding that the (optional) Frequencies
table ("frequencies.txt") does not seem to function at all when used
in a GTFS feed. At first, I used the schedule-testing application
(which is downloadable from Google Transit) on sample test data with a
Frequencies table, and I found that only the absolute times in
stop_times.txt are reflected in the results. The Frequencies table
seems to be completely ignored. I initially thought I must be doing
something wrong or that there might be a bug in the schedule-testing
application, but I've gone back and reviewed the sample live test feed
of "Demo Transit Authority" in Beatty, Nevada that is linked off of
the GTFS specification page on Google, and I'm noticing the same
effect to be true here. (This demo is at:
http://www.google.com/transit?ttype=dep&saddr=North+Ave+at+N+A+Ave+Beatty,+NV&daddr=W+Cottonwood+Dr+at+A+Ave+S+Beatty,+NV&ie=UTF8
.) In this unit test, the fictional bus route being used is "CITY",
the trip on this route is "CITY1" (the opposite-direction trip is
"CITY2"), the source stop is "NANAA", and the destination stop is
"EMSI". When trying to plot the route I find that no matter what
desired day and time I enter for "Depart at", it only shows a
departure from "NANAA" at the following occurrence of 6:06 a.m. and an
arrival at "EMSI" at 6:34 a.m. This is roughly consistent with the
absolute times in stop_times.txt, but it seems to violate the
intervals in
frequencies.txt, which has entries for periods throughout the entire
day from 6:00:00 to 22:00:00 (10 p.m.). If I try to search for a
route at noon, for example, it will return that the next available
trip is at 6:06 a.m. the next day, when it should really return
something on the same day.
Is the functionality of the Frequencies table in GTFS still in beta or
having any known bugs? I have browsed some of the publicly-accessible
GTFS databases from PublicFeeds, and it appears that almost all
agencies don't use the Frequencies table except for a few entries with
very long-term periods (i.e. a whole day), and rather brute-force/
generate the list of times in stop_times.txt.
Am I doing something wrong?