bm/dbt-precompiled-model-hash #26
Conversation
bmoore813
had a problem deploying
to
Approve Integration Tests
GitHub Actions
Failure
bmoore813
had a problem deploying
to
Approve Integration Tests
GitHub Actions
Failure
| {%- do comment_dict.update({ | ||
| "dbt_raw_code_hash": local_md5(node.raw_code) | ||
| }) -%} | ||
| {%- endif -%} |
There was a problem hiding this comment.
@bmoore813 just curious, does raw_code refer to the raw SQL with the jinja references + model config block at the top?
There was a problem hiding this comment.
I love the question and happy to test further but it was my understanding that depending on the node in jinja it could be an option of these so if it exists then it pulls from the graph and applies what the raw_code is....How do you think I could better test to ensure this is the raw dbt code that is in question
There was a problem hiding this comment.
but yes I would assume this is what the code looks like on the repo
There was a problem hiding this comment.
I just looked at what the raw_code looks like in the manifest.json and it is indeed exactly as I just suggested.
NiallRees
had a problem deploying
to
Approve Integration Tests
GitHub Actions
Failure
This will allow tagging to automatically have a tag that uses the hash of the precompiled dbt model so they can better measure the delta in cost when a change is made to a model