Hey all, anyone encountered DataBricks Hive ingest...
# ingestion
b
Hey all, anyone encountered DataBricks Hive ingestion error
ValueError: ('# Detailed Table Information', None, None) is not in list
? I encounter this for all tables, but all database names are ingested fine.
l
@careful-pilot-86309 ^
h
Hey @billowy-book-26360 can you try using the recipe below:
Copy code
source:
    type: hive
    config:
        host_port: '<http://dbc-xxxx.cloud.databricks.com:443|dbc-xxxx.cloud.databricks.com:443>'
        scheme: databricks+connector
        options:
            connect_args:
                http_path: <http-path>
                access_token: <access-token>
Make sure you have sqlalchemy-databricks and databricks-sql-connector installed. It worked for me. Let me know if you run into issues.
b
Thanks @hundreds-photographer-13496 that seems to be working! I don't see detailed table info like Type, Provider, Location etc. in the UI. How would I see if that was being ingested with this connector?
h
Thats great! Unfortunately, detailed table info is not ingested when using
scheme : databricks+connector
The fix for the ValueError you mentioned earlier will be available soon in acryl-pyhive python package . Once that's available, you can go back to using recipe
scheme: databricks+pyhive
as mentioned in hive docs where tables along with detailed table info will be ingested. I will update here once it's available. cc: @dazzling-judge-80093
b
Thanks @hundreds-photographer-13496, that's great. I'll look out for the update. Strangely I no longer get the ValueError using
databricks+pyhive
after a few reboots and some additional unrelated pip installs.
h
Interesting! Do you see detailed table info like Type, Provider, Location etc ?
b
No, I re-ran
databricks+connector
ingestion but Properties tab under Hive objects (tables) is blank. When I use
databricks+connector
I see is_view=True under Properties tab and view_definition is blank.