Null-safe equality for unique_key within delete+insert and merge incremental strategies

[amardatar]

Null-safe equality for unique_key within delete+insert and merge incremental strategies with an appropriate null comparison.

resolves #159

Problem

The current implementation of the incremental merge does not use null-safe equality, meaning if any of the columns are null, duplicate values will be inserted every time the model is run. This is unlikely to be desired behaviour for anyone.

Solution

The SQL for the delete condition is updated to handle a check for null values.

Checklist

• I have read the contributing guide and understand what's expected of me
• 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
• This PR has no interface changes (e.g. macros, cli, logs, json artifacts, config files, adapter interface, etc) or this PR has already received feedback and approval from Product or DX

[github-actions]

Thank you for your pull request! We could not find a changelog entry for this change. For details on how to document a change, see the contributing guide.

[nikita-sheremet-flocktory]

@dbeatty10

I appologies that reach you directly but is there news for this PR? If there is additional help or actions required? Could you please tell me wether help is needed and if I can I willl try to help. This PR fixed the issue that is really annoying :(

Thanks in advance!

[nikita-sheremet-flocktory]

@amardatar

First of all many thanks for such usefull PR! You provided changes for default__get_delete_insert_merge_sql that implements delete+insert strategy. But there is also additional function default__get_merge_sql that responsible for merge strategy and it has it's own {% if unique_key %}.

Could you please add or is null condition for it also?

Thanks in advance!

[peterallenwebb]

@colin-rogers-dbt, I Reviewed this and thought about it with @dbeatty10 a bit. We think it looks good.

[dbeatty10]

@peterallenwebb and I took a closer look at this today, and we think there's a couple more test cases to add:

1. Check null comparison when the unique_key has a single column (also with multiple rows with null values)
2. Check null comparison when the unique_key has a multiple columns (also with multiple rows with null values)

I will take responsibility for trying out these cases and adding these tests.

[dbeatty10]

Important

After decoupling dbt-adapters from dbt-core, only the unit tests run in CI; none of the functional tests run in CI in this repo.

So we'll need to run the changes in this PR along within the functional tests in dbt-postgres (and possibly also dbt-snowflake, dbt-bigquery, and dbt-redshift) in order to know that the changes work as intended.

[rightx2]

Can't wait this feature!