[Github connector] Hi team. Is there a reason why ...
# replication-troubleshooting
r
[Github connector] Hi team. Is there a reason why there is no support for incremental replication for the
pull_request_commits
stream? Is it a technical limitation with GH's api?
m
I think it can be add a incremental sync mode for this endpoint, probably using the same strategy as Commits and other streams: https://github.com/airbytehq/airbyte/blob/c472340ad34a0c61c440ea22256c6b854abd61f9[…]-integrations/connectors/source-github/source_github/streams.py
Not completly sure, but you can test it.
n
Let me know if Marcos' comment helps, Renato!
r
Hi, thanks for the reply, happy to try it out and contribute a PR if it works, I'll keep you posted on this thread.
n
Thanks, looking forward to your PR and update!
r
Hi @Nataly Merezhuk (Airbyte). Looks like it is indeed a technical limitation. Airbyte's incremental system relies on some sort of timestamp and this specific stream does not have any
created_at
or
updated_at
. The only other way I could think about solving this would be to only fetch
pull_request_commits
for pull-requests created or changed after the last sync but not really sure this is even possible within Airbyte's architecture. From my understanding, streams are fetched independently of each other.
n
Hi Renato, I looked into the schema for that stream and there's a date field. I think this field could be used to setup an incremental sync. I don't have time to test this out today, but will later in the week. If you have a chance to test this though, let me know!
e
I would love to have the incremental sync as Renato's idea! Just created a discussion on Github.