When does Airbyte create tables in the destination, and how does it decide the schema?

[rahilsondhi]

Airbyte created this table in Snowflake using the wrong schema. created_at and updated_at should be timestamps, but they're varchar. Also, why did it cluster it by _airbyte_emitted_at? It should not be manually clustered.

When I click on Source > Fields, it shows the correct schema:

So I'm just confused why Airbyte created the table that way.

I tried dropping the table in Snowflake, and I triggered a new sync in Airbyte, hoping that it would re-create the table with the correct schema, but it did not, even though the sync "succeeded."

I wanted to try the Reset your data button, but the docs say it will delete all the rows in my destination tables. Is that in my entire Snowflake database, or just the active streams within this connection? Because I only have one stream/table in this connection, the webhooks table. It's fine if Airbyte truncates that table, but I'm terrified that it will truncate all tables in the Snowflake database.

---

To summarize, my questions are:

1. Does Airbyte create tables in the destination?
   a. If so, when does it do it? Does it verify the table exists before starting a sync?
   b. How does it come up with the schema? Why did it pick varchar for timestamp columns?
   c. How does it pick the cluster by?
2. I deleted the Snowflake table and triggered an Airbyte sync, and it succeeded, but the table doesn't exist, so what happened?
3. Can we get more specific about what Reset your data does? I hope it only truncates the active tables for the current connection, not all tables in the destination

Thank you!