feat: make sure that device_manufacturer is lowered prior to joining

sql_generators/mobile_kpi_support_metrics/templates/active_users.view.sql

@@ -3,10 +3,7 @@ CREATE OR REPLACE VIEW
   `{{ project_id }}.{{ dataset }}.{{ name }}`
 AS
 SELECT
-  * EXCEPT (isp) REPLACE(
-    -- Lower device_manufacturer as in some cases the same manufacturer value has different casing.
-    LOWER(device_manufacturer) AS device_manufacturer
-  ),
+  * EXCEPT (isp),
   CASE
     WHEN LOWER(isp) = "browserstack"
       THEN CONCAT("{{ friendly_name }}", " ", isp)
@@ -60,5 +57,8 @@ SELECT
     ELSE
       CAST(NULL AS STRING)
   END AS device_type,
+  -- Lower device_manufacturer as in some cases the same manufacturer value has different casing.
+  -- But keeping the original value to not create a sudden impact on downstream calculations.
+  LOWER(device_manufacturer) AS device_manufacturer_lowered,
 FROM
   `{{ project_id }}.{{ dataset }}.baseline_clients_last_seen`

sql_generators/mobile_kpi_support_metrics/templates/engagement.query.sql

@@ -2,8 +2,8 @@
 WITH device_manufacturer_counts AS (
   SELECT
     submission_date,
-    device_manufacturer,
-    DENSE_RANK() OVER(PARTITION BY submission_date ORDER BY COUNT(*) DESC) AS manufacturer_rank,
+    device_manufacturer_lowered,
+    RANK() OVER(PARTITION BY submission_date ORDER BY COUNT(*) DESC) AS manufacturer_rank,
   FROM
   `{{ project_id }}.{{ dataset }}.engagement_clients`
   WHERE
@@ -16,7 +16,7 @@ WITH device_manufacturer_counts AS (
     {% endraw %}
   GROUP BY
     submission_date,
-    device_manufacturer
+    device_manufacturer_lowered
 )
 
 SELECT
@@ -36,12 +36,12 @@ SELECT
   COUNTIF(is_mau) AS mau,
   device_type,
   -- Bucket device manufacturers with low count prior to aggregation
-  IF(manufacturer_rank <= 150, device_manufacturer, "other") AS device_manufacturer,
+  IF(manufacturer_rank <= 150, device_manufacturer_lowered, "other") AS device_manufacturer,
 FROM
   `{{ project_id }}.{{ dataset }}.engagement_clients`
 LEFT JOIN
   device_manufacturer_counts
-  USING(submission_date, device_manufacturer)
+  USING(submission_date, device_manufacturer_lowered)
 WHERE
   {% raw %}
   {% if is_init() %}

sql_generators/mobile_kpi_support_metrics/templates/engagement_clients.view.sql

@@ -19,7 +19,7 @@ WITH active_users AS (
     is_mau,
     is_mobile,
     device_type,
-    device_manufacturer,
+    device_manufacturer_lowered,
   FROM
     `{{ project_id }}.{{ dataset }}.active_users`
 ),
@@ -65,7 +65,7 @@ SELECT
     ELSE 'Unknown'
   END AS lifecycle_stage,
   device_type,
-  device_manufacturer,
+  device_manufacturer_lowered,
 FROM
   active_users
 LEFT JOIN

sql_generators/mobile_kpi_support_metrics/templates/new_profile_clients.view.sql

@@ -14,7 +14,7 @@ SELECT
   normalized_os AS os,
   normalized_os_version AS os_version,
   device_model,
-  device_manufacturer,
+  device_manufacturer_lowered,
   is_mobile,
   {% for attribution_field in product_attribution_fields %}
   attribution.{{ attribution_field }},

sql_generators/mobile_kpi_support_metrics/templates/new_profiles.query.sql

@@ -2,8 +2,8 @@
 WITH device_manufacturer_counts AS (
   SELECT
     first_seen_date,
-    device_manufacturer,
-    DENSE_RANK() OVER(PARTITION BY submission_date ORDER BY COUNT(*) DESC) AS manufacturer_rank,
+    device_manufacturer_lowered,
+    RANK() OVER(PARTITION BY first_seen_date ORDER BY COUNT(*) DESC) AS manufacturer_rank,
   FROM
   `{{ project_id }}.{{ dataset }}.new_profile_clients`
   WHERE
@@ -16,7 +16,7 @@ WITH device_manufacturer_counts AS (
     {% endraw %}
   GROUP BY
     first_seen_date,
-    device_manufacturer
+    device_manufacturer_lowered
 )
 
 SELECT
@@ -29,7 +29,7 @@ SELECT
   os,
   os_version,
   -- Bucket device manufacturers with low count prior to aggregation
-  IF(manufacturer_rank <= 150, device_manufacturer, "other") AS device_manufacturer,
+  IF(manufacturer_rank <= 150, device_manufacturer_lowered, "other") AS device_manufacturer,
   is_mobile,
   {% for field in product_attribution_fields.values() if not field.client_only %}
   {{ field.name }},
@@ -40,7 +40,7 @@ FROM
   `{{ project_id }}.{{ dataset }}.new_profile_clients`
 LEFT JOIN
   device_manufacturer_counts
-  USING(first_seen_date, device_manufacturer)
+  USING(first_seen_date, device_manufacturer_lowered)
 WHERE
   {% raw %}
   {% if is_init() %}

sql_generators/mobile_kpi_support_metrics/templates/retention.query.sql

@@ -2,8 +2,8 @@
 WITH device_manufacturer_counts AS (
   SELECT
     submission_date,
-    device_manufacturer,
-    DENSE_RANK() OVER(PARTITION BY submission_date ORDER BY COUNT(*) DESC) AS manufacturer_rank,
+    device_manufacturer_lowered,
+    RANK() OVER(PARTITION BY submission_date ORDER BY COUNT(*) DESC) AS manufacturer_rank,
   FROM
   `{{ project_id }}.{{ dataset }}.retention_clients`
   WHERE
@@ -18,7 +18,7 @@ WITH device_manufacturer_counts AS (
     {% endraw %}
   GROUP BY
     submission_date,
-    device_manufacturer
+    device_manufacturer_lowered
 )
 
 SELECT
@@ -42,12 +42,12 @@ SELECT
   COUNTIF(repeat_profile) AS repeat_profiles,
   device_type,
   -- Bucket device manufacturers with low count prior to aggregation
-  IF(manufacturer_rank <= 150, device_manufacturer, "other") AS device_manufacturer,
+  IF(manufacturer_rank <= 150, device_manufacturer_lowered, "other") AS device_manufacturer,
 FROM
   `{{ project_id }}.{{ dataset }}.retention_clients`
 LEFT JOIN
   device_manufacturer_counts
-  USING(submission_date, device_manufacturer)
+  USING(submission_date, device_manufacturer_lowered)
 WHERE
   {% raw %}
   {% if is_init() %}

sql_generators/mobile_kpi_support_metrics/templates/retention_clients.view.sql

@@ -15,7 +15,7 @@ WITH active_users AS (
     days_active_bits,
     is_mobile,
     device_type,
-    device_manufacturer,
+    device_manufacturer_lowered,
   FROM
     `{{ project_id }}.{{ dataset }}.active_users`
 ),
@@ -84,7 +84,7 @@ SELECT
     ELSE 'Unknown'
   END AS lifecycle_stage,
   active_users.device_type,
-  active_users.device_manufacturer,
+  active_users.device_manufacturer_lowered,
 FROM
   `{{ project_id }}.{{ dataset }}.baseline_clients_daily` AS clients_daily
 INNER JOIN