Add back depends_on for seeds - only macros, never nodes#6851
Conversation
gshank
left a comment
There was a problem hiding this comment.
I think the approach is fine. We do already use getattr in a couple of other places and it's the only real place to put a user friendly error message if somebody tries to use the missing and prohibited methods.
|
+1 on this please |
|
It occurred to me that instead of using setattr, we could provide "property" methods for ref/sources/metrics that throws an error. Either way is fine with me. |
@gshank I like that much better! Significantly less magic. |
* Extend functional tests for seeds w hooks * Add MacroDependsOn to seeds, raise exception for other deps * Add changelog entry * Fix unit tests * Update upgrade_seed_content * Cleanup * Regen manifest v8 schema. Fix tests * Be less magical * PR feedback (cherry picked from commit 298bf8a)
…, never `nodes` (#6920) * Add back `depends_on` for seeds - only `macros`, never `nodes` (#6851) * Extend functional tests for seeds w hooks * Add MacroDependsOn to seeds, raise exception for other deps * Add changelog entry * Fix unit tests * Update upgrade_seed_content * Cleanup * Regen manifest v8 schema. Fix tests * Be less magical * PR feedback (cherry picked from commit 298bf8a) * Update manifest v8 for v1.4.x * Cleanup --------- Co-authored-by: Jeremy Cohen <jeremy@dbtlabs.com>
|
Hi Team, Which release would the changes in this PR make it into? |
|
@sfc-gh-pkommini This has been backported for inclusion in the next v1.4.x patch release. There are a few more in-progress bug fixes we're hoping to include; at this point, we'll probably put out a release candidate next week. |
|
Second - looking for an update as we're blocked from upgrading until this bug is included in a release. |
|
@sfc-gh-pkommini @smitsrr This fix has been included in a release candidate of v1.4.2 (next patch). Could you try it out and confirm that it fixes the issue for you? |
resolves #6806
Description
depends_onI'm not thrilled with this approach, which requires a bit of mucking with magic methods, but it does keep the changes fairly self-contained.
If we do move forward with this approach, we'll need to:
1.4.latestChecklist
I have opened an issue to add/update docs, or docs changes are not required/relevant for this PRchangie newto create a changelog entry