fix: enable schema evolution tests for s3 datalake

[subodh1810]

Issue: https://github.com/airbytehq/airbyte-internal-issues/issues/11225

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
airbyte-docs	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Jan 23, 2025 8:55am

[edgao]

what's the difference between the base test and the copied one? at the very least, we should have a comment explaining that

(and depending on what the difference is, we maybe should add an option to the base test to switch between these behaviors, instead of copying the method wholesale into this one destination)

[subodh1810]

what's the difference between the base test and the copied one? at the very least, we should have a comment explaining that

(and depending on what the difference is, we maybe should add an option to the base test to switch between these behaviors, instead of copying the method wholesale into this one destination)

There are 2 differences

1. The base test is also changing the data type of a given column from int to string. That conversion is not allowed in Iceberg. It will throw an exception
2. The order of elements expected. The order that base test expects and the order in which iceberg returns are different. The data is exact same, just the order is different.

[edgao]

ah, makes sense. I'll write a ticket to make the test more flexible; in the meantime, can you add that info to the javadoc comment? just so it's easy for us to find later on

[edgao]

(https://github.com/airbytehq/airbyte-internal-issues/issues/11389)