[Bug] Unclear error message when a column in a contracted model is missing a `data_type` about dbt-redshift

[dbeatty10]

Reprex

Create project files as described in dbt-labs/dbt-core#8063.

Then run:

dbt seed && dbt run -s my_model 

And get an error like this:

23:28:27  Unhandled error while executing 
'data_type'
23:28:27  1 of 1 ERROR creating sql table model dbt_dbeatty_tools.my_model ............... [ERROR in 0.07s]
23:28:28  
23:28:28  Finished running 1 table model in 0 hours 0 minutes and 3.88 seconds (3.88s).
23:28:28  
23:28:28  Completed with 1 error and 0 warnings:
23:28:28  
23:28:28    'data_type'
23:28:28  
23:28:28  Done. PASS=0 WARN=0 ERROR=1 SKIP=0 TOTAL=1

There's a full stack trace within logs/dbt.log, but it wasn't clear to me exactly where the true issue was.

Proposed solution

To match the order in dbt-postgres, dbt-bigquery, and dbt-snowflake, just flip the order of these two lines:

dbt-redshift/dbt/include/redshift/macros/adapters.sql

Lines 51 to 52 in f95c534

	{{ get_table_columns_and_constraints() }}
	{{ get_assert_columns_equivalent(sql) }}

It's supposed to be this before this:

    {{ get_assert_columns_equivalent(sql) }}
    {{ get_table_columns_and_constraints() }}

Nice-to-have

Add some defensive code immediately above here to raise a helpful error message when the data_type key is not defined:

            if not v.get("data_type"):
                raise ColumnTypeMissingError([col_name])