it is very handy to create a shared data connection and store login/password etc. but I do not want to use glue catalog or crawler. Is it possible to use data connection just to get connection info/credentials for glue spark job and then use spark methods, bypassing Glue catalog and dynamic frames and such?
found much better way that does not require using boto3 and hardcoding region name:
glueContext.extract_jdbc_conf('my_database_connection_name')
returns something like that:
{'enforceSSL': 'true', 'skipCustomJDBCCertValidation': 'false', 'url': 'jdbc:sqlserver://xxxxxx', 'customJDBCCertString': '', 'user': 'xxxx', 'customJDBCCert': '', 'password': 'xxxxxxxxxx', 'vendor': 'sqlserver'}