Cross-db incremental loads not working in Snowflake

[joshpeng]

Issue

Cross-db incremental loads aren't qualified correctly in Snowflake

Issue description

On Snowflake, when you override the target database of a model in dbt_project.yml like this:

# in dbt_project.yml
extra:
      database: ldw
      schema: extra

A --full-refresh works great for making the initial table into the desired database even when that database is a different one than the one of the target profile.

However, it errors out when doing an incremental load.

Results

This is the stack trace of the error

2019-05-09 16:28:41,774 (Thread-1): On xspend: BEGIN
2019-05-09 16:28:41,940 (Thread-1): SQL status: SUCCESS 1 in 0.17 seconds
2019-05-09 16:28:41,940 (Thread-1): Using snowflake connection "xspend".
2019-05-09 16:28:41,940 (Thread-1): On xspend: use schema extra;
2019-05-09 16:28:42,283 (Thread-1): Snowflake error: 002043 (02000): 018c1420-0274-2253-0000-5a0100302056: SQL compilation error:
Object does not exist, or operation cannot be performed.
2019-05-09 16:28:42,283 (Thread-1): On xspend: ROLLBACK

Notice the use schema extra; command. It should have been a fully qualified use schema ldw.extra;

System information

dbt 0.13.0
macOS
python 3.6.7

Steps to reproduce

1. target profile database = abc
2. dbt_project.yml, set a model's database to be def
3. run the model as an incremental load
4. dbt errors out due to non-fully qualified object

[drewbanin]

Thanks @joshpeng! I am pretty sure the bug you encountered will be fixed in 0.14.0, as we'll be using a merge statement for incremental models instead of a temp table + delete + insert. The point remains that the create_table macro implementation on Snowflake isn't database-aware, and we should fix that too :)