Remove user_id from dimension and add it as "metric" (#81)

In essence we are going to treat user_id the same way we treat domain_sessionid_array, as we may have more than one value and this will cause unwanted extra rows.

models/media_ad_views/scratch/snowplow_media_player_media_ad_views_this_run.sql

@@ -40,7 +40,6 @@ events_this_run as (
     ,ev.user_identifier
     ,ev.session_identifier
     ,ev.app_id 
-    ,ev.user_id
     ,ev.play_id
     ,ev.media_ad_break__break_id as ad_break_id
     ,ev.media_ad__ad_id as ad_id
@@ -58,7 +57,8 @@ events_this_run as (
       {%- endif -%}
       {% endfor -%}
     {%- endif %}
-
+
+    ,max(ev.user_id) as user_id
     ,max(ev.media_label) as media_label
     ,max(ev.media_ad_break__name) as ad_break_name
     ,max(ev.media_ad_break__break_type) as ad_break_type
@@ -81,8 +81,8 @@ events_this_run as (
     ,{{ snowplow_utils.get_string_agg('original_session_identifier', 'ev', is_distinct=True) }} as domain_sessionid_array
 
   from events_this_run as ev
+  {{ dbt_utils.group_by(n=9+(var('snowplow__ad_views_passthroughs', [])|length)) }}
 
-  {{ dbt_utils.group_by(n=10+(var('snowplow__ad_views_passthroughs', [])|length)) }}
 
 )