Add column comments to Snowflake views

[spencer-taylor-workrise]

resolves #17

Description

Adds functionality to persist column descriptions to Snowflake views upon creation.
Expands column name testing (upper/lowercase) for the column-level description integration tests.

Checklist

• I have signed the CLA
• I have run this code in development and it appears to resolve the stated issue
• This PR includes tests, or tests are not required/relevant for this PR
• I have updated the CHANGELOG.md and added information about my change to the "dbt-snowflake next" section.

[jtcohen6]

Closing and re-opening to trigger adapter integration tests, now that I've added ok to test label

Edit: There are some unrelated test failures, related to recent changes to dbt-core. We'll work to get those sorted shortly

[jtcohen6]

Nice work so far on this @spencer-taylor-workrise! Thanks for proving out the approach. A few comments from me, thinking through best ways to keep the logic organized

[jtcohen6]

1. Let's avoid running this query if not config.persist_column_docs(), most likely by nesting it within the existing conditional below
2. The get_columns_in_query macro already handles adding where false limit 0 + wrapping in a subquery

Suggested change

	{% set query_columns = get_columns_in_query(sql + ' limit 0') %}
	{% set query_columns = get_columns_in_query(sql) %}

[spencer-taylor-workrise]

Updated in 66312f7.

[jtcohen6]

Nice work sketching this out! I think we should have two goals:

• avoid duplicating logic between here + snowflake__alter_column_comment down below
• keep the create_view_as macro clean + legible

I've got some pseudo-code below with ideas of how to do it. Feel free to take/leave/adapt as you see fit.

First, we can wrap the common bits of comment generation in a new macro:

{% macro get_column_comment_sql(column_name, column_dict) %}
    {{ adapter.quote(column_name) if column_dict[column_name]['quote'] else column_name }} COMMENT $${{ column_dict[column_name]['description'] | replace('$', '[$]') }}$$
{% endmacro %}

Then, create_view_as can call a separate macro if we're persisting column-level docs:

{% if config.persist_column_docs() -%}
    {% set model_columns = model.columns %}
    {% set query_columns = get_columns_in_query(sql) %}
    {{ get_persist_docs_column_list(model_columns, query_columns) }}
{% endif %} 

That macro would look like:

{% macro get_persist_docs_column_list(model_columns, query_columns) %}
    {% for column_name in query_columns %}
        {% if (column_name|upper in model_columns) or (column_name in model_columns) %}
          {{ get_column_comment_sql(column_name, model_columns) }}  
        {% else %}
          {{ column_name }}
        {% endif %}
        {{ ", " if not loop.last else "" }}
      {% endfor %}
{% endmacro %}

And then the other macro down below:

{% macro snowflake__alter_column_comment(relation, column_dict) -%}
    {% set existing_columns = adapter.get_columns_in_relation(relation) | map(attribute="name") | list %}
    alter {{ relation.type }} {{ relation }} alter
    {% for column_name in column_dict if (column_name in existing_columns) or (column_name|upper in existing_columns) %}
        {{ get_column_comment_sql(column_name, column_dict) }} {{ ',' if not loop.last else ';' }}
    {% endfor %}
{% endmacro %}

[spencer-taylor-workrise]

Updated in 66312f7.

[jtcohen6]

Do we need to get existing_columns?

[spencer-taylor-workrise]

Nope, I think this is left over from a previous version I was testing out. Removing this shortly.

[spencer-taylor-workrise]

Hey @jtcohen6 ! Thanks for the comments -- completely agree with the logic there. These have all been resolved in the most recent commit.

[jtcohen6]

@spencer-taylor-workrise Nice work, on a quick turnaround! This looks good to me :)