Revert "Support for redshift 821" #10446
Conversation
This reverts commit 6e4564a.
|
Thank you for your pull request! We could not find a changelog entry for this change. For details on how to document a change, see the contributing guide. |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #10446 +/- ##
==========================================
- Coverage 88.81% 88.81% -0.01%
==========================================
Files 180 180
Lines 22573 22562 -11
==========================================
- Hits 20049 20038 -11
Misses 2524 2524
Flags with carried forward coverage won't be shown. Click here to find out more.
|
| raise ParsingError( | ||
| "Unit Test fixtures require at least one row free of Nones to ensure consistent column types." | ||
| ) |
There was a problem hiding this comment.
I'd suggest isolating this exception only to adapters that will actually fail to execute if this is the case to make this backward-compatible in terms of expected behaviour.
For example, we could introduce a validate_row(row, row_idx) macro and call it here in the base implementation: https://github.com/dbt-labs/dbt-adapters/blob/782a32b530e25afe34af6896392d9fb081dce9d1/dbt/include/global_project/macros/unit_test_sql/get_fixture_sql.sql#L26-L27
By default, the macro would simply pass but adapters could implement the macro to raise an exception (or warning if they wouldn't fail 100% of the time) if the row.
Just proposing a high-level approach here, don't meant to be overly-prescriptive about the suggested interfaces (e.g. maybe validate_rows (plural) makes more sense instead of passing an index along).
Reverts #10366
Since this will introduce parsing failure to projects that currently do not fail.