Välkommen till Trafiklab:s användare- och supportforum! Ställ frågor, rapportera problem och hjälp oss med förslag och idéer!
Vid felrapporter ber vi dig inkludera exakt API-namn och om möjligt ett exempelanrop för att underlätta felsökningen. Glöm inte att ta bort din API-nyckel när du delar ditt exempelanrop.
Undrar du hur du får tillgäng Trafiklabs data? Läs vår introduktion här: https://www.trafiklab.se/hur-gor-jag
Welcome to Trafiklab's user and supportforum. Ask questions, report issues and help us improve with suggestions and ideas!
If you open a new issue, please always include the exact API you're talking about, and, if applicable, include a sample request so we can check if contains the right parameters. Don't forget to remove your API keys when sharing example requests.
Välkommen till Trafiklab:s kund- och supportforum! Ställ frågor, rapportera problem och hjälp oss med förslag och idéer!
Did you replace {api_key} with your api key when downloading the file?
Regards,
Bert
Yes I did. The download was successful but the file itself is corrupted.
Luke
Which browser did you use to download the file?
The browser may not have stored the file correctly, for example by not handling the compression during transport correctly. During transport, the file is compressed using gzip, so if the browser did not handle this correctly the file may still be gzip compressed. Renaming it to skane.zip.gz may resolve the issue in this case.
Another possibility is an incomplete download. The total file size for todays file should be 33,4mb, 35 082 444 bytes.
Unfortunately this still doesn't help the issue, I tried using safari and google chrome with the same result. Even when adding .gz to the file name the result is the same.
The last GTFS update where this wasn't an issue was Mon 10/02/2025, 07:30, can you see if there was some format change after this update?
I have re-run our data export to recreate the same file again, in order to ensure the problem doesn't lie there. Otherwise I'd expect it to be some issue during file transfer, of you're sure nothing changed on your side.
Regards,
Bert
I've attached the last GTFS that works (file name starts with dk-) and also the most recent file (skane.zip). The GTFS that works came through our automatic pipeline for context but the latter I downloaded directly from the URL. I thought this may help the investigation.
Let me know if there's anything else I can do to help.
Kind regards
Both files you uploaded open normally when I download them. The date for which your import stopped working coincides with the date we added a new file to our datasets, booking_rules.txt. It might be worth to verify if it's really the download which is failing in your automated pipeline, and not the import of this new file booking_rules.
We have not heard any other complaints regarding the ingestion of these zip files, which leads me to believe that the zip files are correctly compressed and should open without issues. Does the old file open on your computer without issues? In that case it may be worth to try opening a recent file downloaded through your automatic pipeline.
Regards,
Bert
We found the issue with the file for our systems, it can read the zip (although I still can't manually) but the issue arrises as an IO error because stop_times has entries which are not referenced in booking_rules (121120000365600738).
Thanks for your help investigating this Bert.
Kind regards
Glad to hear you found the issue.
Please provide the full stop_time row and dataset in case there are rows in stop_times referring to a non-existing booking_rule, so we can fix any possible bugs in our datasets. I looked up the id you provided 121120000365600738, but it matches a row in booking_rules:
stop_times.txt
booking_rules.txt
Regards,
Bert
It seems like we have to investigate further how we want to handle booking_rules as it's causing issues at various points in our pipeline.
No need to investigate any further, we can manually remove this file and any references to make it fit.
Thanks for your help again Bert, sorry if I lead you down a rabbit hole!
If you did not receive an email regarding this upcoming change in january, please check if your email address on developer.trafiklab.se is correct and whether the email may have been classed as spam.
Regards,
Bert