Ft Add materialization for Snowflake materialized view object type

[leahtessem]

Resolves #1162
This is an issue opened for dbt-core repository, but because the code I have written is specific to (and only tested on Snowflake), I feel that it fits in this repository. This PR (or parts of it) can be adapted and moved to live in dbt-core if desired.

Description

• Adding materialization for Snowflake materialized views
  • including support for cluster_by config entry
  • Adding SnowflakeRelationType with materialized view type (including space)
  • add methods to SnowflakeRelation to determine if relation is MV
  • add get relation type method that includes new MV addition
• Update snowflake__list_relations_without_caching macro to include materialized views
  • MVs do not return from show terse objects so it now has to be two calls (one for tables & one for views)
• Also added macro for create or replace-ing materialized view similar to how the base view macro was built out in dbt-core

Handling switching between view types

• Add handling in view materialization to drop the object if a materialized view exists with the desired view name (similar to how tables are handled in the create_or_replace_view macro in dbt-core)
• Add onto incremental materialization to drop the MV if one exists with the desired incremental table name

Added MVs in the switching relation type tests as well as their own test file

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.

[leahtessem]

Hi @dbt-labs/oss-maintainers, hoping to get a review here. I would like feedback on whether any of these changes should go to dbt-core instead and making sure that my tests are sufficient.

Getting this change merged in would allow my organization to have all of our snowflake objects managed in one place. Currently, our materialized views are managed in a way that is disjoint and prone to failure. It would also address issue dbt-labs/dbt-core#1162 and add the feature that has been desired since 2018.

Thanks!

[jtcohen6]

@leahtessem Thanks for the PR, and sorry for the delay!

I owe you a response here. To make a long story short, I think we're unlikely to merge this into dbt-core / adapters for native support — at least for now. For the last few years, we've kept "experimental" implementations of materialized views in this repo: https://github.com/dbt-labs/dbt-labs-experimental-features/tree/main/materialized-views

I'd like to say more about that, though. It's always worth reexamining our long-standing assumptions. I'll plan to swing back in the next few days with a fuller explanation.

[Fleid]

Closing as my current thinking is that we won't support Snowflake's Materialized Views, but Dynamic Tables instead.

Thanks a lot for your contribution, it will inspire our work going forward on that topic.