[Sqlalchemy-tickets] Issue #3810: Engine creation issue when attempting to use an Azure Data Warehouse DB using mssql+pyodbc (zzzeek/sqlalchemy)

[Felipe <iss...@bi...>]

New issue 3810: Engine creation issue when attempting to use an Azure Data Warehouse DB using mssql+pyodbc
https://bitbucket.org/zzzeek/sqlalchemy/issues/3810/engine-creation-issue-when-attempting-to

Felipe:

Whenever trying to create an engine object using the mssql dialect and pyodbc driver, an error will be raised indicating the database_principal_id() does not exist.

According to Microsoft Documentation, this is true for their data warehouse, but is still supported in normal Azure Databases. (https://msdn.microsoft.com/en-us/library/ms187319.aspx)

With my limited knowledge, I applied a workaround but I am not completely sure the impact it might have in other connections.

Edited the **base.py** module to have the following:

Line 532:

```
#!python

   MS_AZURE_VERSION = (12,)
```


Line 1654:
    
```
#!python

def _get_default_schema_name(self, connection):
        if self.server_version_info < MS_2005_VERSION:
            return self.schema_name
        elif self.server_version_info >= MS_AZURE_VERSION:
            query = sql.text("""
                SELECT user_name()
            """)
        else:
            query = sql.text("""
                SELECT default_schema_name FROM
                sys.database_principals
                WHERE principal_id=database_principal_id()
            """)
        default_schema_name = connection.scalar(query)
        if default_schema_name is not None:
            return util.text_type(default_schema_name)
        else:
            return self.schema_name
```