# ask-community

Akira Renbokoji

12/01/2022, 2:03 AM
Is there a way to pass Dagster Redshift connection to pandas read_sql() function? I saw that there was a function __get__conn() but I want to respect its privacy (naming convention). The reason, if it helps to show my purpose of the use is that before I had a connection setup that I was passing into
df = pd.read_sql(some_query, my_redshift_connection)
I would then get back a DataFrame with column names. Now using
my_data = context.resources.redshift.execute_query(some_query, fetch_results=True)
I have to do something like
df = pd.DataFrame(my_data, columns=['my_first_column', 'second_column', 'third'])
I am looking for a way not to declare column names and just get it off my query or DataFrame creation.

Zach P

12/01/2022, 2:14 AM
I'm not familiar with the dagster redshift connection, but it's entirely possible to create your own connection and define it as a resource. You potentially wouldn't even need to write your own code really, just dig into the dagster redshift resource, i wouldn't be surprised if there's something already there but I'm unsure if this is supported by pd.read_sql. TLDR: Yes, but i haven't done this exactly so i can't say if it's easy or may require a few extra lines of code.

Akira Renbokoji

12/02/2022, 5:32 AM
I'm confused. Your response sounds like the exact setup that I have done and not what I'm wanting to do.


12/02/2022, 6:04 PM
I feel like it's pretty reasonable for us to allow you to key into the connection itself, this is likely just an oversight. Mind filing an issue?

Akira Renbokoji

12/02/2022, 6:10 PM