INVALID_PARAMETER_VALUE: Cannot update tag 'mlflow.rootRunId' when importing model into Databricks

[amesar]

When importing into Databricks a registered mode that was exported from Databricks, get the following error: INVALID_PARAMETER_VALUE: Cannot update tag 'mlflow.rootRunId'

[amesar]

Fix: https://github.com/amesar/mlflow-export-import/commit/24b9efb4dd03e1a2c89f2caa140834475ea9612b

[wqp89324]

@amesar , I'm testing the two databricks notebooks, namely Export_Experiment and Import_Experiment, by first exporting and then importing within one databricks account. The export part worked fine by exporting to a dbfs directory. However, when I try importing from the dbfs directory, I got the INVALID_PARAMETER_VALUE: Cannot update tag 'mlflow.rootRunId' error. The peek_at_experiment seems to work fine. I also tried setting import_metadata_tags to True or False, but getting the same error. Could you please help? I suspect you made the changes above for model and run, but not for experiment?

[wqp89324]

Update: I get the same error when trying Import_Model, and Export_Model seems to work well.

[wqp89324]

@amesar could you please reopen this issue?

[hebo-yang]

+1

[amesar]

@wqp89324 and @hebo-yang. Just checked in a fix. There was an edge case for nested run tags when importing into Databricks. Let me know if it works.