I’m ingesting data from Snowflake. During the POC we ran into some duplicate events getting loaded due to not having the right warehouse size combined with retrying loading. I believe it was mentioned that if we provided a unique identifier for an event during ingestion we could prevent that in the future if any retries occurred. I don't see an option for how to provide an event id to amplitude in the ingestion documentation though. Any ideas on how to add that in?
Best answer by Saish Redkar
You’ll have to send an insert_id field for each record. You can find the additional supported fields in the HTTP docs - https://www.docs.developers.amplitude.com/analytics/apis/http-v2-api/#keys-for-the-event-argument
+10You’ll have to send an insert_id field for each record. You can find the additional supported fields in the HTTP docs - https://www.docs.developers.amplitude.com/analytics/apis/http-v2-api/#keys-for-the-event-argument
Thanks
Just to validate. Does this insert_id also applies for the snowflake ingestion? We are not using the HTTP v2 API
Thanks!
+10Yes, their Snowflake ingestion is based on the HTTP API underneath.
We send the insert_id along with the other requisite fields/columns in our Snowflake Ingest config SQL.
If you don't have an Amplitude account, you can create an Amplitude Starter account for free and enjoy direct access to the Community via SSO. Create an Amplitude account. You can also create a Guest account below!
If you don't have an Amplitude account, you can create an Amplitude Starter account for free and enjoy direct access to the Community via SSO. Create an Amplitude account. Want to sign up as a guest? Create a Community account.
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.
