❄️ New in

dagster-snowflake

0.18.0 ❄️ This release contains a breaking change for the Snowflake Resource. If you only use the Snowflake I/O manager this will not affect you. Previously, the behavior of the

execute_query

and

execute_queries

methods of the Snowflake Resource handled the

fetch_results

and

use_pandas_result

differently. This release brings the two methods in line with each other, however that requires a breaking change. Here’s a table outlining how these parameters affect the returns of the

execute_queries

and

execute_query

methods

fetch_results | use_pandas_result | outcome/return type
     True     |       False       | list   (fetchall() used)
     True     |       True        | pd.DataFrame (list of pd.DataFrames for execute_queries) (fetch_pandas_all() used)
     False    |       True        | Exception raised
     False    |       False       | None

You can find more information on return types in the Snowflake docs (we use the

fetchall

and

fetch_pandas_all

methods as described in the table) Additionally, in previous versions, if you set

use_pandas_result=True

in

execute_queries

we would attempt to append the DataFrame result of each query into a single DataFrame. Now the DataFrame results of each query are returned as a list of DataFrames.