Hello, so I am trying to reingest and reprofile fr...
# troubleshoot
m
Hello, so I am trying to reingest and reprofile from Hive, but I keep getting this error why I don't understand why appears:
Copy code
"AttributeError: 'bytes' object has no attribute 'value'\n
I suppose that the problem might be related to a problem of the parse process during the porfiling (as if the profile option is disabled, this error does not appear), but I don't know why I am getting it (the last successful ingestion was on Friday and during the weekend no one has touched the deployment). I will share the full log in the next message of the thread I would appreciate any help to narrow down the source of the error!
h
It looks like this issue might be due to great-expectations version 0.15.23 released 3 days ago. Datahub uses great-expectations for profiling. For now, I would suggest using Datahub CLI with
great-expectations>=0.15.12, <0.15.23
, t verify if it appears there as well.
m
The strange thing is that I have other sources that are profiled when executed and when the ingestion is executed this error does not appear. I tested in previously this morning in the CLI, but I will try it again and I will be back with the results.
So apparently via CLI worked but I am not 100% sure on why it worked. The thing is that if I download the hive plugin (as it is also done in the UI internally)
pip install 'acryl-datahub[hive]'
the ingestion will end up with the previously shown error. But if such plugin isn't installed (which in previous versions of the actions framework led to an error as it wasn't installed by default), the ingestion works correctly. I am guessing that, apart from installing the dependencies needed for the Hive ingestion to work, the newer version of great-expectations that you have mentioned is also installed, which leads to obtaining this error.
The image I am using on the actions framework is the following
acryldata/datahub-actions:2aad384