search

databricks-demos / dbdemos

Demos to implement your Databricks Lakehouse
Other
278 stars 89 forks source link

lakehouse360 churn demo: DLT errors out when populating the churn_prediction table #141

Open tomidatabricks opened 2 weeks ago

tomidatabricks commented 2 weeks ago

Summary: In the lakehouse360 churn demo: DLT errors out when populating the churn_prediction table with the error: python_model = cloudpickle.load(f) ^^^^^^^^^^^^^^^^^^^ TypeError: code() argument 13 must be str, not int

Setting the TLD pipeline to current channel like resolved the issue.

here is a slack discussion regarding this: https://databricks.slack.com/archives/C02GZJSLAR0/p1726523349629669

Trying to run the lakehouse360 churn demo and getting this error in the churn_prediction table of the DLT pipeline: File "/local_disk0/.ephemeral_nfs/envs/pythonEnv-59230cd2-3b21-4d60-ac2f-b7f5d3590a5b/lib/python3.11/site-packages/mlflow/pyfunc/model.py", line 533, in _load_context_model_and_signature python_model = cloudpickle.load(f) ^^^^^^^^^^^^^^^^^^^ TypeError: code() argument 13 must be str, not int

michaelk-dbx commented 1 week ago

I faced today the same issue with the IOT demo, moving to current channel resolved the issue.