From b2989b1e6a1fccabefc15367f8dd25b8e12c48f7 Mon Sep 17 00:00:00 2001
From: Ilias Xenogiannis <ilias1111@users.noreply.github.com>
Date: Thu, 28 Nov 2024 18:12:21 +0200
Subject: [PATCH] Test and fix

---
 .../.scripts/integration_tests.sh             | 17 +++++--
 .../assert_media_ad_views_passthrough.sql     | 45 +++++++++++++++++++
 integration_tests/tests/schema.yml            | 20 +++++++++
 ...w_media_player_media_ad_views_this_run.sql |  4 ++
 4 files changed, 83 insertions(+), 3 deletions(-)
 create mode 100644 integration_tests/tests/assert_media_ad_views_passthrough.sql
 create mode 100644 integration_tests/tests/schema.yml

diff --git a/integration_tests/.scripts/integration_tests.sh b/integration_tests/.scripts/integration_tests.sh
index 1add026..eea8ab0 100755
--- a/integration_tests/.scripts/integration_tests.sh
+++ b/integration_tests/.scripts/integration_tests.sh
@@ -45,8 +45,6 @@ for db in ${DATABASES[@]}; do
 
   eval "dbt run --target $db --vars '{snowplow__enable_youtube: false, snowplow__enable_whatwg_media: false, snowplow__enable_whatwg_video: false, snowplow__enable_media_player_v1: false}'" || exit 1;
 
-
-
   echo "Snowplow media player integration tests: Execute models - run 1/6"
 
   eval "dbt run --target $db --full-refresh --vars '{snowplow__allow_refresh: true}'" || exit 1;
@@ -69,6 +67,19 @@ for db in ${DATABASES[@]}; do
 
   eval "dbt test --target $db" || exit 1;
 
+  echo "Snowplow media player integration tests: Testing ad views passthrough - mixed configuration"
+  eval "dbt run --target $db --full-refresh --vars '{
+    snowplow__allow_refresh: true,
+    snowplow__enable_media_ad: true,
+    snowplow__backfill_limit_days: 999,
+    snowplow__ad_views_passthroughs: [
+      \"v_collector\",
+      {\"sql\": \"v_tracker || app_id\", \"alias\": \"tracker_app_id\", \"agg\": \"max\"}
+    ]
+  }'" || exit 1;
+
+  eval "dbt test --target $db --select test_type:singular tag:ad_views_passthrough" || exit 1;
+
   echo "Snowplow media player integration tests: All tests passed"
 
-done
+done
\ No newline at end of file
diff --git a/integration_tests/tests/assert_media_ad_views_passthrough.sql b/integration_tests/tests/assert_media_ad_views_passthrough.sql
new file mode 100644
index 0000000..9618551
--- /dev/null
+++ b/integration_tests/tests/assert_media_ad_views_passthrough.sql
@@ -0,0 +1,45 @@
+-- tests/assert_media_ad_views_passthrough.sql
+with base_data as (
+  select * from {{ ref('snowplow_media_player_base_events_this_run') }}
+  where media_ad__ad_id is not null 
+    and media_identifier is not null
+),
+
+expected as (
+  select
+    {{ dbt_utils.generate_surrogate_key(['platform', 'media_identifier', 'media_ad__ad_id']) }} as media_ad_id,
+    -- simple passthrough
+    v_collector,
+    -- max aggregation
+    max(v_tracker || app_id) as tracker_app_id
+  from base_data
+  group by 1, 2
+),
+
+actual as (
+  select 
+    media_ad_id,
+    v_collector,
+    tracker_app_id
+  from {{ ref('snowplow_media_player_media_ad_views_this_run') }}
+),
+
+comparison as (
+  select 
+    coalesce(e.media_ad_id, a.media_ad_id) as media_ad_id,
+    case 
+      when e.v_collector != a.v_collector then 1
+      when e.tracker_app_id != a.tracker_app_id then 1
+      when e.media_ad_id is null then 1
+      when a.media_ad_id is null then 1
+      else 0
+    end as has_mismatch
+  from expected e
+  full outer join actual a using (media_ad_id)
+)
+
+select
+  'Data mismatch found' as test_name,
+  sum(has_mismatch) as failures
+from comparison
+having sum(has_mismatch) > 0
\ No newline at end of file
diff --git a/integration_tests/tests/schema.yml b/integration_tests/tests/schema.yml
new file mode 100644
index 0000000..9f9b996
--- /dev/null
+++ b/integration_tests/tests/schema.yml
@@ -0,0 +1,20 @@
+version: 2
+
+models:
+  - name: snowplow_media_player_media_ad_views_this_run
+    tests:
+      - assert_media_ad_views_passthrough
+    columns:
+      - name: media_ad_id
+        tests:
+          - not_null
+          - unique
+      - name: v_collector
+        tests:
+          - not_null
+      - name: tracker_app_id
+        tests:
+          - not_null
+
+    config:
+      tags: ['this_run']
\ No newline at end of file
diff --git a/models/media_ad_views/scratch/snowplow_media_player_media_ad_views_this_run.sql b/models/media_ad_views/scratch/snowplow_media_player_media_ad_views_this_run.sql
index 0a12632..ae4b8ca 100644
--- a/models/media_ad_views/scratch/snowplow_media_player_media_ad_views_this_run.sql
+++ b/models/media_ad_views/scratch/snowplow_media_player_media_ad_views_this_run.sql
@@ -145,6 +145,10 @@ select
     {%- for col in passthrough_names %}
       , p.{{col}}
     {%- endfor -%}
+    {%- for agg_col in agg_columns %}
+      , p.{{agg_col['alias']}}
+    {%- endfor -%}
+
   {%- endif %}
 
 from prep as p