We currently use trip_id and start_date to identify specific trips. As @derhuerst pointed out here and in a private conversation, this is not enough to identify journeys in frequency-based trips. This is also clarified in the GTFS-RT spec.
This issue has low priority because (to our current knowledge) no service in Germany has frequency-based trips.
Anyway, we need to check if this issue also affects "Verstärkerfahren", that is, additional trips that are not found in the schule but resemble scheduled trips and might re-use their trip_id.
We currently use
trip_id
andstart_date
to identify specific trips. As @derhuerst pointed out here and in a private conversation, this is not enough to identify journeys in frequency-based trips. This is also clarified in the GTFS-RT spec.We need to:
start_time
column to our db_schema hereupdate_statement
andinsert_statement
inper_schedule_importer
start_time
is guaranteed to contain datastart_time
to match rows in curves.rsThis issue has low priority because (to our current knowledge) no service in Germany has frequency-based trips.
Anyway, we need to check if this issue also affects "Verstärkerfahren", that is, additional trips that are not found in the schule but resemble scheduled trips and might re-use their trip_id.