TBH the initial fix did feel slightly wonky, and now that you mention it I'm 
starting to view this whole ordeal in a new light. You know what they say about 
the proverbial frog in boiling water..

It is my understanding that the currently identified problematic SQLA 
databases/connectors (Snowflake, Redshift, Oracle) share the following 
characteristics:
- `cursor.description` returns all lowercase column names.
- Executing a select via the cursor with lowercase column names (or mixed/upper 
case for that matter) results in upper case column labels in the query result.

While it can't be excluded that some engine down the line will exhibit some 
other behaviour, I suspect having a normalizing function when creating a tables 
that uppercases column names when returned from a `cursor.description` should 
solve most of the observed problems. In addition, metrics would probably need 
to be passed through the same normalizer, but beyond that one might not need to 
do any further wrangling (famous last words).

I think I'll give it a shot and see where I land.

[ Full content available at: 
https://github.com/apache/incubator-superset/pull/5686 ]
This message was relayed via gitbox.apache.org for [email protected]

Reply via email to