dstandish commented on pull request #13152: URL: https://github.com/apache/airflow/pull/13152#issuecomment-753166870
Let me caveat that I am not a committer and can't merge this anyway, so my opinion is not decisive. I just took a look cus I have experience with snowflake and can help review. And let me also say that I totally get the impulse to add this -- as someone who has used snowflake for a couple years I also was annoyed by the weird handling of case. And also, let me share that I know what it's like to try to contribute or propose something and get a "meh". That all having been said, I just don't think this should be added. It's just too niche, and it's not worth the complexity and confusion it adds. True it's not a _lot_ of complexity, but it's another configurable thing, and it is nonstandard, not part of snowflake api, not part of pandas api, and not supported in other hook methods. I think it is best to simply accept snowflake as it is. `get_pandas_df` returns the naming that snowflake gives it. If they are lower and quoted in the system, they'll be lower in pandas. And the same is true with fetchall or querying from an IDE or the web UI for that matter. Snowflake currently has a parameter you can set to make quoted identifiers case insensitive. And when enabled, even quoted identifiers are rendered as upper case. I think the best solution would be if snowflake extended this to make it possible that quoted identifiers are case insensitive _and_ always rendered as lower. But that's an issue for another team :) Anyway, just my 2 cents. Others very well may feel differently. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
