[ISSUE #35110] match CATs records only one primary key when primary k…

[maxi297]

…ey is defined

What

Addresses airbytehq/airbyte-internal-issues#6376

How

If source_defined_primary_key is defined, validate only primary key values. Handle the both case exact_order = true and exact_order = false

Test exact_order = False (stripe)

CATs are currently failing for stripe because subscription_items, subscription_schedule, subscriptions and usage_records are now empty so in order to test, I did:

• add subscription_items, subscription_schedule, subscriptions and usage_records as empty streams
• change txn_1KVQhfEcXtiJtvvhF7ox3YEm to txn_1KQhfEcXtiJtvvhF7ox3YEm for balance_transactions
• ran python -m pytest integration_tests -p integration_tests.acceptance -k TestBasicRead

Outcome:

Note: the stream name was added since the printscreen was taken

Test exact_order = True (google-ads)

In order to test, I did:

• inverse the order of the first two ad_group_ad_legacy expected records
• ran python -m pytest integration_tests -p integration_tests.acceptance -k TestBasicRead

Outcome:

Note: a specific assertion message has been added since the printscreen was taken

If someone wants more validation, I can change one of the primary key fields value or any other tests.

🚨 User Impact 🚨

This is a relaxation of CATs. Hence, we expect more CATs to start passing and CATs to be less flaky. The drawback for now is is that we will lack visibility on new fields that are not added on the root level (root level is validated by Datadog) as NoAdditionalPropertiesValidator will not be used anymore until we have a better plan for schema validation.

[vercel]

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

1 Ignored Deployment

Name	Status	Preview	Comments	Updated (UTC)
airbyte-docs	⬜️ Ignored (Inspect)	Visit Preview		Feb 27, 2024 6:24pm

[alafanechere]

minor suggestions, I'll let you decide if they're worth it. LGTM 👍 tried out on #35587 : it's working.