Migration guide for Attribution

Upgrading to 0.6.0

• We had to introduce limitations for Redshift users:
  1. users are only allowed to have 1 path transformation defined within the snowplow__path_transforms variable
  2. the snowplow__path_transforms variable parameters only accept single-item arrays for remove_if_last_and_not_all and remove_if_not_all
• This is due to Redshift deprecating Python UDFs; the transformations now rely on native SQL, which allows less flexibility for automation.
• Existing users can stay on older versions of the package as long as their UDFs remain in the database.

ymldbt_project.yml

vars:
  snowplow_attribution:
    snowplow__path_transforms: {'exposure_path': null}

Upgrading to 0.5.0

• From now on the snowplow__path_transforms variable parameters only accept non-empty arrays for remove_if_last_and_not_all and remove_if_not_all variables instead of strings, please your variable overwrites in your dbt_project.yml accordingly. Previously you could only remove one specific channel or campaign, now you can do multiple, if needed.

ymldbt_project.yml

vars:
  snowplow_attribution:
    snowplow__path_transforms: {'exposure_path': null, 'remove_if_last_and_not_all': ['channel_to_remove_1', 'campaign_to_remove_1, 'campaign_to_remove_2']}