We have some disparity in explore table names between streams and datasets - for streams we add the "cdap_stream_" prefix, while for datasets we use the dataset name as is.
We could have a conflict if someone created a dataset with the name cdap_stream_<existing-stream-id> in the same namespace. Unlikely, but possible. Also not consistent between streams and datasets.
Also, since the hive database that we create is already in the format cdap_[namespace-id], we do not need the cdap_ prefix in the table name again.
Although the fix may be simple, opening a separate bug to track this, since there could be some upgrade implications if this were fixed.
The namespace-id in the hive table name is also redundant because the table is already in a hive database that is named cdap_[namespace-id]
, closing this since the fix went as in as part of your upgrade explore PR, and we decided to keep the 'stream_' and 'dataset_' prefixes.