Snapshotting with check_cols='all' doesn't handle schema changes when new columns are added

[joshpeng]

Describe the bug

When the source table has schema change via a new column addition, dbt doesn't handle the change when using check_cols='all'. The check_cols is likely trying to hash the non-existent column in the snapshot table before adding the column to the table.

Steps To Reproduce

1. Snapshot a source table with strategy='check' and check_cols='all'
2. Add a column to the source table
3. Run dbt snapshot

Expected behavior

dbt to add the new column and continue along snapshotting gracefully.

Screenshots and log output

Database Error in snapshot pallets_snapshot (snapshots/snapshot_teal_pallets.sql)
  000904 (42000): 018f405d-0215-21e7-0000-5a0102fe42da: SQL compilation error: error line 206 at position 12
  invalid identifier 'SNAPSHOTTED_DATA.PALLET_TYPE_ID'

System information

Which database are you using dbt with?

• snowflake

The output of dbt --version:

installed version: 0.14.2
   latest version: 0.14.2

Up to date!

The operating system you're using:
macOS

The output of python --version:
Python 3.7.3